of used subcarriers 1728

Frame duration, *Tframe* 5ms DL/UL rate 2:1 OFDM symbols in the DL 30 Number of transmit antennas, *M* {1,2,4} Number of receive antennas, *N* {1,2,4} MIMO detector MMSE *gd=*(*N* – *M* + 1)=1. At higher data rates (*R*>8), all the codes perform similarly in the analysed SNR range despite of the different diversity order between them.

#### **4.4.4 TACS performance under bit error rate minimization criterion**

In Fig. 7 and Fig. 8, the bit error rate performance using TACS is shown having a fixed rate *R*=4. Fig. 7 shows the improvement due to the increase in *Ma* and also the performance achieved when combined with code selection. It can be observed how the TAS increases the diversity order, leading to a large performance increase for the SM and Golden subsets. It is very important to notice that despite the diversity increase for all the LDC subsets, SD and SIMO schemes still perform better when each code is evaluated independently. However, in Fig. 8, we can observe that when the code selection is switched on, SIMO and Golden subsets are selected most times, while the usage of SIMO increases with the SNR and the usage of SM and the Golden code increases with *Ma*. Furthermore, the achieved improvement by the TACS is clearly appreciated in Fig. 7, where an SNR improvement of approximately 1dB is obtained for *Ma*={3,4}. It is also surprising that the SM code is rarely selected knowing that the Golden code should always outperform SM since it obtains a higher diversity. However, as it is observed in Fig. 8, for less than 5% of the channel realizations the SM may outperform slightly the Golden code. Whether the singular value decomposition of the effective channel H is analysed when SM is selected, it has been observed that when all singular values are very close, both the SM and the Golden code lead to very similar performances, therefore no matter which one is selected.

In Fig. 9 and Fig. 10, the performance using the TACS is again analysed for *R*=8. In Fig. 9 the different diversity orders of SD, SM, and the Golden Code are illustrated. We can appreciate here that the SM and the Golden code show the best performance when *Ma*={3,4}, and also for *Ma*=2 when SNR≤18dB. Furthermore the increase in the diversity order due to TACS can be observed in both Fig. 7 and Fig. 9. The maximum diversity order (*gd* = *MaN*) is achieved since at least one LDC (SIMO and G2) from those in the codebook are able to achieve the maximum diversity order.

Moreover, the BER using the TACS is equivalent to that obtained from the SISO scheme (referred as SISO*eq* in the plots) over a Rayleigh fading channel with the same rate *R,* a diversity order *gd*=*MaN* and a coding gain equal to . The performance of this *equivalent* SISO scheme, in terms of the bit error rate probability *Pb*, can be obtained directly by close expressions that are found in [41][42] and applying the Craig's formula in [43],

$$P\_b = \frac{1}{\log\_2 \sqrt{Z}} \sum\_{i=1}^{\log\_2 \sqrt{Z}} P\_b(i) \tag{28}$$

$$P\_b(\mathbf{i}) = \frac{2}{\sqrt{Z}} \sum\_{k=0}^{\left(1-2^{-\frac{r}{\alpha}}\right)\sqrt{Z}-1} \alpha(k, i, Z) \frac{1}{\pi} \int\_0^{\pi/2} \left(1 + \left(2k + 1\right)^2 \frac{3\rho\_b}{2\left(Z - 1\right)\sin^2\theta}\right)^{-\frac{\rho\_d}{\alpha}} d\theta \tag{29}$$

$$\log(k, i, Z) = (-1) \left| \begin{array}{c} \frac{k \cdot 2^{i-1}}{\sqrt{Z}} \left| \left( 2^{i-1} - \left\lfloor \frac{k \cdot 2^{i-1}}{\sqrt{Z}} + \frac{1}{2} \right. \right. \right. \end{array} \right| \right. \tag{30}$$

Space-Time Adaptation and MIMO Standardization Status 119

Fig. 7. Uncoded BER performance when *N*=2, *R=*4, *Ma*={2,3,4} for uncorrelated MIMO

Fig. 8. LDC selection statistics with *N*=2, *R=*4, *Ma*={2,3,4} for uncorrelated MIMO Rayleigh

Rayleigh channel and MMSE linear receiver.

channel and MMSE linear receiver.

where *<sup>b</sup>* = · / *log*2(Z), *x* means the smallest integer of *x*, and Z is the modulation order of the *Z*-QAM modulation.

The values of for different combinations of *Ma*={2,3,4}, *N*={2,3,4} and *R*={4,8} are depicted in Table 2. These values have been obtained adjusting the BER approximation in Eq. (28) to the empirical BER. As shown in Fig. 7 and Fig. 9 the performance of the TACS schemes is perfectly parameterized under the equivalent SISO model. Notice also that the power gain is constant across the whole SNR range.


Table 2. Coding gain for the TACS proposal with *Ma*={2,3,4}, *N*={2,3,4}, and *R*={4,8}.

Fig. 6. Uncoded BER for uncorrelated Rayleigh channel with MMSE detector and *N*=2.

The values of for different combinations of *Ma*={2,3,4}, *N*={2,3,4} and *R*={4,8} are depicted in Table 2. These values have been obtained adjusting the BER approximation in Eq. (28) to the empirical BER. As shown in Fig. 7 and Fig. 9 the performance of the TACS schemes is perfectly parameterized under the equivalent SISO model. Notice also that the power gain is

*Ma*=2 2.66 3.9 6.31

*Ma*=3 3.20 5.2 8.41

*Ma*=4 3.75 6.2 9.44

*Ma*=2 4.20 9 14

*Ma*=3 6.75 14.5 23

*Ma*=4 9.00 19 28.5

Table 2. Coding gain for the TACS proposal with *Ma*={2,3,4}, *N*={2,3,4}, and *R*={4,8}.

Fig. 6. Uncoded BER for uncorrelated Rayleigh channel with MMSE detector and *N*=2.

*N*=2 N=3 N=4

/ *log*2(Z), *x* means the smallest integer of *x*, and Z is the modulation order

where *<sup>b</sup>* = ·

R = 4

R = 8

of the *Z*-QAM modulation.

constant across the whole SNR range.

Fig. 7. Uncoded BER performance when *N*=2, *R=*4, *Ma*={2,3,4} for uncorrelated MIMO Rayleigh channel and MMSE linear receiver.

Fig. 8. LDC selection statistics with *N*=2, *R=*4, *Ma*={2,3,4} for uncorrelated MIMO Rayleigh channel and MMSE linear receiver.

Space-Time Adaptation and MIMO Standardization Status 121

In this section the performance of the TACS adaptation scheme in case the throughput is maximized (see Eq.(27)) is analysed. Then, for such adaptation scheme, the antenna set and the LDC code that maximizes the throughput is selected. In addition, the highest MCS (in the sense of spectral efficiency) that achieves a BLER<0.01 (1%) is also selected. The look-uptable used for mapping the ESINR to the BLER is shown and described in [14]. In the scenarios considered, the minimum allocable block length according the IEEE 802.16e standard was selected [17] (i.e. the number of sub-channels *Nsch* occupied per block varies

In Fig. 11 and Fig. 12, the spectral efficiency achieved by TACS with adaptive Modulation and Coding (AMC) as well as the LDC statistics are shown. For Spatial Multiplexing (SM), two encoding options named *Vertical Encoding* (VE) and *Horizontal Encoding* (HE) are considered. For the first scheme, VE, the symbols within the codeword apply the same MCS format, whereas for the second, HE, each symbol may apply a different MCS. Clearly the first is more restrictive since is limited by the worst stream (*min*(*ESNRq*)) whereas the second is able to exploit inter-stream diversity at the expense of higher signalling requirements (at

Depicted performances shown that at low SNRs (SNR<13dB), the SIMO and Alamouti achieve the highest spectral efficiencies (something that has been already obtained in several previous works [10]). However, as the SNR is increased, the codes with higher multiplexing capacity (e.g. the SM and the Golden code) are preferred. It could be also observed that the SM with VE implies a loss of around 2dB compared to the Golden code, but when HE is

Fig. 11. Spectral efficiency under TACS with throughput maximization criterion with *Ma*=2, *N*=2, adaptive MCS and MMSE receiver for an uncorrelated MIMO Rayleigh channel.

**4.4.5 TACS performance under throughput maximization criterion** 

between 1 and 4). The number of available antennas is *Ma*=2 whereas *N*=2.

least twice as that required with VE in case of *M*=2).

used, the Golden code is around 0.5dB worse than the SM-HE.

Fig. 9. Uncoded BER performance when *N*=2, *R=*8, *Ma*={2,3,4} for uncorrelated MIMO Rayleigh channel and MMSE linear receiver.

Fig. 10. LDC selection statistics when *N*=2, *R=*8, *Ma*={2,3,4} for uncorrelated MIMO Rayleigh channel and MMSE linear receiver.

Fig. 9. Uncoded BER performance when *N*=2, *R=*8, *Ma*={2,3,4} for uncorrelated MIMO

Fig. 10. LDC selection statistics when *N*=2, *R=*8, *Ma*={2,3,4} for uncorrelated MIMO Rayleigh

Rayleigh channel and MMSE linear receiver.

channel and MMSE linear receiver.

#### **4.4.5 TACS performance under throughput maximization criterion**

In this section the performance of the TACS adaptation scheme in case the throughput is maximized (see Eq.(27)) is analysed. Then, for such adaptation scheme, the antenna set and the LDC code that maximizes the throughput is selected. In addition, the highest MCS (in the sense of spectral efficiency) that achieves a BLER<0.01 (1%) is also selected. The look-uptable used for mapping the ESINR to the BLER is shown and described in [14]. In the scenarios considered, the minimum allocable block length according the IEEE 802.16e standard was selected [17] (i.e. the number of sub-channels *Nsch* occupied per block varies between 1 and 4). The number of available antennas is *Ma*=2 whereas *N*=2.

In Fig. 11 and Fig. 12, the spectral efficiency achieved by TACS with adaptive Modulation and Coding (AMC) as well as the LDC statistics are shown. For Spatial Multiplexing (SM), two encoding options named *Vertical Encoding* (VE) and *Horizontal Encoding* (HE) are considered. For the first scheme, VE, the symbols within the codeword apply the same MCS format, whereas for the second, HE, each symbol may apply a different MCS. Clearly the first is more restrictive since is limited by the worst stream (*min*(*ESNRq*)) whereas the second is able to exploit inter-stream diversity at the expense of higher signalling requirements (at least twice as that required with VE in case of *M*=2).

Depicted performances shown that at low SNRs (SNR<13dB), the SIMO and Alamouti achieve the highest spectral efficiencies (something that has been already obtained in several previous works [10]). However, as the SNR is increased, the codes with higher multiplexing capacity (e.g. the SM and the Golden code) are preferred. It could be also observed that the SM with VE implies a loss of around 2dB compared to the Golden code, but when HE is used, the Golden code is around 0.5dB worse than the SM-HE.

Fig. 11. Spectral efficiency under TACS with throughput maximization criterion with *Ma*=2, *N*=2, adaptive MCS and MMSE receiver for an uncorrelated MIMO Rayleigh channel.

Space-Time Adaptation and MIMO Standardization Status 123

different directions, and in most cases a trade-off between both is meet by each specific space-time code. From a system point of view, and due to the inherent time/freq variability of the wireless channel, no code is optimal for all channel conditions, and at most, the codes can be optimized according to the ergodic properties of the channel. In fact, this is the reason why the TACS scheme is able to bring significant gain compared to a scheme where the same space-time code is always used. This situation is well-known and it is the reason why in most of the Broadband Wireless Access (BWA) systems, the number of space-time

In IEEE 802.16e/m, two types of MIMO are defined, Single User MIMO and Multiuser MIMO, the first corresponding to the case where one resource unit (the minimum block of frequency-time allocable subcarriers) is assigned to a single user, and the second when this

In case of two transmit antennas, IEEE 802.16e/m defines two possible encoding schemes referred as Matrix A and Matrix B. Matrix A corresponds to the Alamouti scheme, while Matrix B corresponds to the Spatial Multiplexing (SM) case. In case of using SM, WiMAX allows both Vertical Encoding (VE) and Horizontal Encoding (HE). In the first case, VE, all the symbols are encoded together and belong to the same *layer*. In addition to Matrix A and Matrix B, IEEE 802.16 also defines a Matrix C which corresponds to the Golden Code. This code is characterized for providing the highest spatial diversity for the spatial rate *R*=2. In case of 3 and 4 transmit antennas, WiMAX also defines the encoding schemes of Matrix A, Matrix B, and Matrix C, all of them providing different trade-offs between diversity and

The list of combinations is even longer since WiMAX allows antenna selection and antenna grouping, therefore, the list of encoding matrices also includes the possibility that not all antennas are used, and only a subset are selected (the list of matrices in Table 3 do not show this possibility). In case not all the antennas are used, the power is normalized so that the

Besides the possibility to select among any of the previous coding matrices, IEEE 802.16e/m also allows the use of precoding. In this case, the space-time coding output is weighted by a

where *x* is *Mt*×1 vector obtained after ST encoding, where Mt is the number of streams at the output of the space time coding scheme. The matrix W is a *M*×*Mt* weighting matrix where M is the number of transmit antennas. The weighting matrix accepts two types of adaptation depending on the rate of update, named short term closed-loop precoding and long term

In the later IEEE 802.16m, the degrees of flexibility has been broadened, allowing several kinds of adaptation [44]. On top of this, IEEE 802.16m includes also ST codes for up to 8 transmitter antennas, enabling the transmission at spectral efficiencies as high as 30bits/sec/Hz which become necessary to achieve the very high throughputs demanded for

*z Wx* (31)

same power is transmitted disregard of the number of active antennas.

matrix before mapping onto transmitter antennas

codes is increasing.

spatial multiplexing.

closed-loop precoding.

IMT-Advanced systems [45].

one is shared among multiple users.

Fig. 12. LDC selection statistics under TACS with throughput maximization criterion with *Ma*=2, *N*=2, adaptive MCS and MMSE receiver for an uncorrelated MIMO Rayleigh channel.

To gain further insights of the TACS behaviour, the statistics of LDC selection as a function of the average SNR are plotted in Fig. 12. We can clearly appreciate that at low SNR the preferred scheme is SIMO where all the power is concentrated in the best antenna, while as the SNR is increased full rate codes (*Q=M*) are more selected since they permit to use lower size constellations. Moreover, comparing SM-VE with SM-HE, we can observe that SM-HE is able to exploit the stream's diversity and hence achieves a higher spectral efficiency than if the Golden code is used. Actually, at average SNR=12, the SM with HE is the scheme selected for most frames, even more than SIMO. These results show that in case of linear receivers (e.g. MMSE) the TACS scheme with AMC gives a noticeable SNR gain (up to 3dB) in a large SNR margin (SNR from 6 to 18dB) and also is a good technique to achieve a smooth transition between diversity and multiplexing.

#### **5. MIMO in IEEE 802.16e/m**

The use of MIMO may improve the performance of the system both in terms of link reliability and throughput. As it was discussed in previous sections, both concepts pull in

Fig. 12. LDC selection statistics under TACS with throughput maximization criterion with *Ma*=2, *N*=2, adaptive MCS and MMSE receiver for an uncorrelated MIMO Rayleigh channel.

To gain further insights of the TACS behaviour, the statistics of LDC selection as a function of the average SNR are plotted in Fig. 12. We can clearly appreciate that at low SNR the preferred scheme is SIMO where all the power is concentrated in the best antenna, while as the SNR is increased full rate codes (*Q=M*) are more selected since they permit to use lower size constellations. Moreover, comparing SM-VE with SM-HE, we can observe that SM-HE is able to exploit the stream's diversity and hence achieves a higher spectral efficiency than if the Golden code is used. Actually, at average SNR=12, the SM with HE is the scheme selected for most frames, even more than SIMO. These results show that in case of linear receivers (e.g. MMSE) the TACS scheme with AMC gives a noticeable SNR gain (up to 3dB) in a large SNR margin (SNR from 6 to 18dB) and also is a good technique to achieve a

The use of MIMO may improve the performance of the system both in terms of link reliability and throughput. As it was discussed in previous sections, both concepts pull in

smooth transition between diversity and multiplexing.

**5. MIMO in IEEE 802.16e/m** 

different directions, and in most cases a trade-off between both is meet by each specific space-time code. From a system point of view, and due to the inherent time/freq variability of the wireless channel, no code is optimal for all channel conditions, and at most, the codes can be optimized according to the ergodic properties of the channel. In fact, this is the reason why the TACS scheme is able to bring significant gain compared to a scheme where the same space-time code is always used. This situation is well-known and it is the reason why in most of the Broadband Wireless Access (BWA) systems, the number of space-time codes is increasing.

In IEEE 802.16e/m, two types of MIMO are defined, Single User MIMO and Multiuser MIMO, the first corresponding to the case where one resource unit (the minimum block of frequency-time allocable subcarriers) is assigned to a single user, and the second when this one is shared among multiple users.

In case of two transmit antennas, IEEE 802.16e/m defines two possible encoding schemes referred as Matrix A and Matrix B. Matrix A corresponds to the Alamouti scheme, while Matrix B corresponds to the Spatial Multiplexing (SM) case. In case of using SM, WiMAX allows both Vertical Encoding (VE) and Horizontal Encoding (HE). In the first case, VE, all the symbols are encoded together and belong to the same *layer*. In addition to Matrix A and Matrix B, IEEE 802.16 also defines a Matrix C which corresponds to the Golden Code. This code is characterized for providing the highest spatial diversity for the spatial rate *R*=2. In case of 3 and 4 transmit antennas, WiMAX also defines the encoding schemes of Matrix A, Matrix B, and Matrix C, all of them providing different trade-offs between diversity and spatial multiplexing.

The list of combinations is even longer since WiMAX allows antenna selection and antenna grouping, therefore, the list of encoding matrices also includes the possibility that not all antennas are used, and only a subset are selected (the list of matrices in Table 3 do not show this possibility). In case not all the antennas are used, the power is normalized so that the same power is transmitted disregard of the number of active antennas.

Besides the possibility to select among any of the previous coding matrices, IEEE 802.16e/m also allows the use of precoding. In this case, the space-time coding output is weighted by a matrix before mapping onto transmitter antennas

$$\mathbf{z} = \mathsf{V}\mathsf{Vx} \tag{31}$$

where *x* is *Mt*×1 vector obtained after ST encoding, where Mt is the number of streams at the output of the space time coding scheme. The matrix W is a *M*×*Mt* weighting matrix where M is the number of transmit antennas. The weighting matrix accepts two types of adaptation depending on the rate of update, named short term closed-loop precoding and long term closed-loop precoding.

In the later IEEE 802.16m, the degrees of flexibility has been broadened, allowing several kinds of adaptation [44]. On top of this, IEEE 802.16m includes also ST codes for up to 8 transmitter antennas, enabling the transmission at spectral efficiencies as high as 30bits/sec/Hz which become necessary to achieve the very high throughputs demanded for IMT-Advanced systems [45].

Space-Time Adaptation and MIMO Standardization Status 125

The use of multiple antenna techniques at transmitter and receiver sides is still considered a hot research topic where the channel capacity can be increased if multiple streams are multiplexed in the spatial domain. The study on the trade-off between diversity and multiplexing has motivated the emergence of many different space-time coding architectures where most of the proposed schemes lie in the form of Linear Dispersion Codes. Furthermore, as it was shown by the authors in previous sections, when the transmitter disposes of partial channel state information, robustness and throughput can be very significantly improved. One of the simplest adaptation techniques is the use of antenna selection, which increases the diversity of the system up to the maximum available (*gd*=*MaNa*). On the other hand, when transmit antenna selection is combined with code selection a coding gain is achieved. In this chapter, a joint Transmit Antenna and space-time Coding Selection (TACS) scheme previously proposed by the authors has been described. The TACS algorithm allows two kind of optimization: *i*) bit error rate minimization, and *ii*) throughput maximization. One important result obtained from these studies is that the number of required space-time coding schemes is quite low. In fact, previous studies by the author have shown that in case of spectral efficiencies of 8bits/second/Hertz or lower, using SIMO, Alamouti, SM, and the Golden code is enough to maximize the performance (for higher rates, codes with higher spatial rate would be required). Furthermore, the worse performance achieved by linear receivers (e.g. ZF, MMSE) is compensated by the TACS scheme, which allows to achieve performances close to those obtained with the non-linear receivers (e.g. the Maximum Likelihood) with much lower computational requirements. As a final conclusion, it can be considered that transmit antenna selection with linear dispersion code selection can be an efficient spatial adaptation technique whose low feedback requirements make it feasible for most of the Broadband Wireless Access systems, especially

**6. Summary** 

in case of low mobility.

BLER Block Error Rate BS Base Station

CSI Channel State Information FDD Frequency Division Duplexing LDC Linear Dispersion Codes LTE Long Term Evolution

MCS Modulation and Coding Scheme MIMO Multiple Input Multiple Output MMSE Minimum Mean Square Error OSTBC Orthogonal Space-Time Block Code QAM Quadrature Amplitude Modulation SIMO Single Input Multiple Output SISO Single Input Single Output SM Spatial Multiplexing SNR Signal To Noise Ratio

3GPP 3rd Generation Partnership Project AWGN Additive White Gaussian Noise

**7. Acronyms** 


Table 3. WiMAX IEEE 802.16e MIMO encoding matrices.

<sup>2</sup> In case of 3 and 4 transmit antennas, Matrix A, B and C accept different antenna grouping and selection schemes. This antenna grouping does similar effects as TACS, indicating which antennas and Space-time codes are preferred.

#### **6. Summary**

124 Advanced Transmission Techniques in WiMAX

Alamouti (a.k.a. Matrix A)

Spatial Multiplexing (a.k.a. Matrix B)

Golden Code (a.k.a. Matrix C)

Matrix A2

Matrix B

Matrix A

Matrix B

M Nmin T Q R MIMO Encoding Matrix Name

*s rs s jrs r*

0 0

*s s*

\* 3 2

2

*s s s s*

\* \* 0 14 5 \* \* 10 54 \* \* 2 36 7 \* \* 32 76

2 In case of 3 and 4 transmit antennas, Matrix A, B and C accept different antenna grouping and selection schemes. This antenna grouping does similar effects as TACS, indicating which antennas and

\* \* 0 14 5 \* \* 10 54 \* \* 6 72 3

*sss* <sup>012</sup> Matrix C

*ssss* <sup>0123</sup> Matrix C

*s ss s ss ss s ss s*

\* \* 10 2 3

 0 3 12 <sup>2</sup> 1 23 0 <sup>1</sup> 1 5 , <sup>2</sup> <sup>1</sup>

*s s s s* \* 0 1 \* 1 0

*s s* 0 1

*s s*

0 0

4

*s s s s*

*ss s s*

\* 0 1

<sup>3</sup> 0 0

 

> <sup>3</sup> 0 0 4 <sup>3</sup> 0 0

*s ss s ss ss s ss s ss ss*

2 2 2 4 2 *s jrs rs s <sup>r</sup>*

2 1 2 2 1

3 2 4 4 1

3 2 4 4 1

4 1 4 4 1

4 2 4 8 2

Space-time codes are preferred.

3 2 4 4 1 *<sup>T</sup>*

4 4 1 4 4 *<sup>T</sup>*

Table 3. WiMAX IEEE 802.16e MIMO encoding matrices.

2 2 1 2 2 *<sup>T</sup>*

The use of multiple antenna techniques at transmitter and receiver sides is still considered a hot research topic where the channel capacity can be increased if multiple streams are multiplexed in the spatial domain. The study on the trade-off between diversity and multiplexing has motivated the emergence of many different space-time coding architectures where most of the proposed schemes lie in the form of Linear Dispersion Codes. Furthermore, as it was shown by the authors in previous sections, when the transmitter disposes of partial channel state information, robustness and throughput can be very significantly improved. One of the simplest adaptation techniques is the use of antenna selection, which increases the diversity of the system up to the maximum available (*gd*=*MaNa*). On the other hand, when transmit antenna selection is combined with code selection a coding gain is achieved. In this chapter, a joint Transmit Antenna and space-time Coding Selection (TACS) scheme previously proposed by the authors has been described. The TACS algorithm allows two kind of optimization: *i*) bit error rate minimization, and *ii*) throughput maximization. One important result obtained from these studies is that the number of required space-time coding schemes is quite low. In fact, previous studies by the author have shown that in case of spectral efficiencies of 8bits/second/Hertz or lower, using SIMO, Alamouti, SM, and the Golden code is enough to maximize the performance (for higher rates, codes with higher spatial rate would be required). Furthermore, the worse performance achieved by linear receivers (e.g. ZF, MMSE) is compensated by the TACS scheme, which allows to achieve performances close to those obtained with the non-linear receivers (e.g. the Maximum Likelihood) with much lower computational requirements. As a final conclusion, it can be considered that transmit antenna selection with linear dispersion code selection can be an efficient spatial adaptation technique whose low feedback requirements make it feasible for most of the Broadband Wireless Access systems, especially in case of low mobility.

#### **7. Acronyms**


Space-Time Adaptation and MIMO Standardization Status 127

[17] IEEE Standard for Local and metropolitan area networks, Part 16: Air Interface for

Licensed Bands and Corrigendum 1, IEEE Std 802.16e™-2005, Feb.2006. [18] M. Vu, A. Paulraj, "MIMO Wireless Linear Precoding", IEEE Signal Processing

[19] A.B. Gershman, N.D. Sidiropoulos, "Space Time Processing for MIMO

[20] R. W. Heath, A.J. Paulraj, "Linear Dispersion Codes for MIMO Systems Based in Frame Theory", IEEE Transactions on Signal Processing, Vol.50, n.10, Oct.2002 [21] R. Gohary, T. Davidson, "Design of Linear Dispersion Codes: Asymptotic Guidelines

[22] R. Heath, S. Sandhu, A. Paulraj, "Antenna Selection for Spatial Multiplexing Systems with Linear Receivers", IEEE Communications Letters, Vol.5, no.4, April 2001. [23] D.A. Gore, A.J. Paulraj, "MIMO antenna subset selection with space-time coding," *IEEE Transactions on Signal Processing,* vol.50, no.10, pp. 2580-2588, Oct 2002. [24] D. Deng, M. Zhao, J. Zhu, "Transmit Antenna Selection for Linear Dispersion Codes

[25] W.C. Freitas, F.R.P. Cavalcanti, R.R. Lopes, "Hybrid MIMO Transceiver Scheme with

[26] R.W. Heath, A.J. Paulraj, "Switching between diversity and multiplexing in MIMO

[27] L. Che, V.V. Veeravalli, "A Limited Feedback Scheme for Linear Dispersion Codes Over

[29] D.J. Love, R.W. Heath, T. Strohmer, "Grassmannian beamforming for multiple-input

[30] D. Deng, J. Zhu, "Linear Dispersion Codes Selection Based on Grassmannian Subspace Packing",submitted to IEEE Journal of Selected Topics in Signal Processing, 2007. [31] R. Machado, B.F. Uchoa-Filho, T.M. Duman, "Linear Dispersion Codes for MIMO

*Conference*, 2008 (WCNC-2008), pp.199-204, March 31 2008-April 3 2008. [32] D. Yang, N. Wu, L.L. Yang, L. Hanzo, "Closed-loop linear dispersion coded eigen-beam

[33] R.W. Heath, D. J. Love, "Multimode Antenna Selection for Spatial Multiplexing Systems

*Signal Processing, 2007*, (ICASSP-2007), vol.3, pp. 41-44, 15-20 April 2007. [28] A. Osseiran, V. Stankovic, E. Jorswieck, T. Wild, M. Fuchs, M. Olsson, "A MIMO

Magazine, Vol. 4, no.5, pp.87-105, Sept. 2007.

Vol.4, No.6, Nov.2005.

Spring.

Sept. 2006.

17-20 June 2007.

September 11 2008.

no.10, pp. 2735-2747, Oct. 2003.

Communications", John Wiley & Sons, UK, 2005.

Fixed and Mobile Broadband Wireless Access Systems, Amendment 2: Physical and Medium Access Control Layers for Combined Fixed and Mobile Operation in

and Their Implementation", IEEE Transactions on Wireless Communications,

Based on Linear Receiver", Proc. Vehicular Technology Conference, 2006. VTC 2006-

Antenna Allocation and Partial CSI at Transmitter Side", *IEEE 17th International Symposium on Personal, Indoor and Mobile Radio Communications, 2006*, pp.1-5, 11-14

systems," *IEEE Transactions on Communications*, vol.53, no.6, pp. 962-968, June 2005.

Correlated MIMO Channels," *IEEE International Conference on Acoustics, Speech and* 

framework for 4G systems: WINNER concept and results", *IEEE 8th Workshop on Signal Processing Advances in Wireless Communications, 2007*, (SPAWC-2007), pp.1-5,

multiple-output wireless systems," *IEEE Transactions on Information Theory*, vol.49,

Channels with Limited Feedback," *IEEE Wireless Communications and Networking* 

transmission and its capacity," *Electronics Letters*, vol.44, no.19, pp.1144-1146,

with Linear Receivers", *IEEE Transactions on Signal Processing*, Vol.53, no.8, Aug.2005.


#### **8. References**


[1] E. Telatar, "Capacity of multi-antenna Gaussian channels", *European Transactions on* 

[2] A. Saad, M. Ismail, N. Misran, "Correlated MIMO Rayleigh Channels: Eigenmodes and

[3] L. Zheng, D. Tse, "Diversity and multiplexing: a fundamental tradeoff in multiple

[4] V. Tarokh, N. Seshadri, A. R. Calderbank, "Space-Time codes for high data rates wireless

[5] V. Tarokh, H. Jafarkhani, A.R. Calderbank, "Space-time block codes from orthogonal designs," *IEEE Transactions on Information Theory,* vol.45, no.5, pp.1456-1467, Jul 1999. [6] G. Ganesan, P. Stoica, "Space-time diversity scheme for wireless communications", in

[7] S. Sandhu, R. Heath, A. Paulraj, "Space-time block codes versus space-time trellis codes,"

[8] J. Cheng; H. Wang; M. Chen; S. Cheng, "Performance comparison and analysis between

[9] L. Yu, P.H.W. Fung, W. Yan, S. Sumei, "Performance analysis of MIMO system with

[10] C. Oestges, B. Clerckx, "MIMO Wireless Communications", Academic Press, Elsevier,

[11] B. Hassibi, B.M. Hochwald, "High Rates Codes that are Linear in Space and Time", April

[12] W. Zhang, X. Ma, B. Gestner, D. V. Andreson, "Designing Low Complexity Equalizers for Wireless Systems", *IEEE Communications Magazine*, January, 2009, pp. 56-62. [13] G. J. Foschini, "Layered space-time architecture for wireless communications in a fading environment when using multiple antennas", Bell Lab Tech. J. v.1., n.2, 1996. [14] I. Gutierrez, "Adaptive Communications for Next Generation Broadband Wireless

[15] S. M. Alamouti, "A simple transmit diversity technique for wireless communications", IEEE J. Selected Areas in Communications, vol. 17, pp. 1451-1458, Oct. 1998. [16] J.C. Belfiore, G. Rekaya, E. Viterbo: "The Golden Code: A 2 x 2 Full-Rate Space-Time

Code with Non-Vanishing Determinants," IEEE Transactions on Information

antenna channels", *IEEE Trans. On Information Theory*, May, 2003.

Information Theory, vol.44, pp.744-765, March, 1998.

Capacity Analyses", *International Journal of Computer Science and Network Security*,

communication: Performance criterion and code construction", IEEE. Trans. on

*IEEE International Conference on Communications*, 2001 (ICC-2001), vol.4, pp.1132-

STTC and STBC," the 54th *IEEE Vehicular Technology Conference*, (VTC 2001 Fall),

serial concatenated bit-interleaved coded modulation and linear dispersion code," *IEEE International Conference on Communications*, 2004, vol.2, no., pp. 692-696 Vol.2,

STBC Space-Time Block Code

TDD Time Division Duplexing UPA Uniform Power Allocation

ZF Zero Forcing

**8. References** 

TACS Transmit Antenna and (space-time) Code Selection

*Telecommunications*, Nov. 1999.

Vol. 8 No. 12 pp. 75-81, Dec. 2008.

vol.4, no., pp.2487-2491 vol.4, 2001.

Access Systems", Ph.D. Thesis, June 2009.

Theory, vol. 51, n. 4, pp. 1432-1436, Apr. 2005.

Proc. ICASSP, 2000.

1136 vol.4, 2001.

20-24 June 2004.

USA, 2007.

2001.


**Part 2** 

**Physical Layer Models and Performance** 


### **Part 2**

### **Physical Layer Models and Performance**

128 Advanced Transmission Techniques in WiMAX

[34] K. Youngwook, C.Tepedelenlioglu, "Threshold-Based Substream Selection for Closed-

[35] R. Machado, B.F. Uchôa-Filho, "Extended Techniques for Transmit Antenna Selection

[36] T. Lestable, M. Jiang, A. Mourad, D. Mazzarese, Uplink MIMO Schemes for IEEE

[37] T. Lestable, M. Jiang, A. Mourad, D. Mazzarese, S. Han, H. Choi, H. Kang , I. Gutierrez,

[38] I. Gutierrez, F. Bader, A. Mourad, Spectral Efficiency Under Transmit Antenna and STC

[39] I. Gutiérrez, F. Bader, A. Mourad, Joint Transmit Antenna and Space-Time Coding

[40] R. Srinivasan et al., "Evaluation Methodology for P802.16m-Advanced Air Interface",

[41] W. Lopes, W. Queiroz, F. Madeiro, M. Alencar, "Exact Bit Error Probability of M-QAM

[42] J. Lu, T. Tjhung, C. Chai, "Error probability Performance of *L*-Branch Diversity

[43] M. K. Simon, "Evaluation of Average Bit Error Probability for Space-Time Coding based

[44] "IEEE Standard for Local and metropolitan area networks Part 16: Air Interface for

[45] Report ITU-R M.2134, Requirements related to technical system performance for

Communications and Networking, 3 (3): 257-267, Sept. 2001.

pp.215-226, Jan. 2008.

C802.16m-08/535, July, 2008.

September 2009, Tokyo (Japan).

IEEE 802.16m-07/037r2.

vol. 46, No.2, Feb. 1998.

REP-M.2134-2008/en>.

Conference, 2007. Oct. 2007.

Std 802.16m-2011, 06-May-2011

802.16m, IEEE S80216m-08\_534r2, July, 2008.

pp.118-195, 2006.

(Qatar).

Loop Spatial Multiplexing", *IEEE Transactions on Vehicular Technology*, vol.57, no.1,

with STBCs", Journal of Communications and Information Systems, vol.21, n.3,

"Linear Dispersion Codes for Uplink MIMO schemes in IEEE 802.16m", IEEE

Selection with Throughput Maximization Using WiMAX, in Proceedings of the 17th International Conference on Telecommunications (ICT2010), 4-7 April 2010, Doha

Selection for WiMAX MIMO System, in Proceedings of the 20th IEEE Personal, Indoor and Mobile Radio Communications Symposium 2009 (PIMRC 2009), 13-16

modulation Over Flat Rayleigh Fading Channels", Microwave and Optoelectronics

Reception of MQAM in Rayleigh Fading", IEEE Transactions on Communications,

on a Simpler Exact Evaluation of Pairwise Error Probability", Journal of

Broadband Wireless Access Systems Amendment 3: Advanced Air Interface", IEEE

IMT-Advanced radio interface(s), November 2008 <http://www.itu.int/publ/R-

**7** 

*Taiwan* 

**Hybrid ARQ Utilizing Lower Rate** 

*Industrial Technology Research Institute (ITRI), Hsinchu,* 

Hybrid automatic repeat request (Hybrid ARQ or HARQ), an extension of ARQ that incorporates forward error correction coding (FEC), is a retransmission scheme with errorcontrol method employed in current communications systems. In standard ARQ, redundant bits are added to data to be transmitted using an error-detecting code, e.g., cyclic redundancy check (CRC). The contribution of HARQ is its efficient utilization of the available resources and the provision of reliable services in latest-generation systems.

This chapter focuses on wireless systems using HARQ with emphasis on the multiple-input multiple-output (MIMO) paradigm. In this chapter, the architecture of MIMO transceivers that are based on bit-interleaved coded modulation (BICM) and it employs HARQ is described. MIMO system is an attractive technique that can enhance the spectral efficiency through spatial multiplexing (SM) [Foschini 1995]. However, many wireless environments may suffer from ill-conditioned channels or multipath fadings, which degrade the system performance. The aim of this chapter is to find an efficient MIMO scheme for retransmission

In IEEE 802.16e [WiMAX 2007] standard, Space Time Coding (STC) subpacket combining, which retransmits with a different MIMO format, has been introduced. One possible way to combine the initial signal and the retransmitted signal is the utilization of Maximum Ratio Combining (MRC) in symbol level based on Virtual Space Time Block Coding (VSTBC) [Gao et. al , 2007]. However, such combining technique works properly only if the channel is quasi-static. In circumstances with high mobility, this technique does not provide satisfactory performance. Another approach is to combine the retransmitted and initial signals using symbol level combining (SLC) before the detector. However, the required buffer in SLC occupies large memory in receivers. Moreover, both SLC receiver architectures are only applicable when the initial and retransmitted symbols are aligned. Thus, they may be impractical for Incremental Redundancy (IR) HARQ [Lin et al. ,1984] and Constellation Rearrangement (CoRe). Although bit level combining (BLC) can solve the problem of huge buffer requirement, using ML-like MIMO detectors, e.g., List Sphere Decoder (LSD) [

to combat the ill-ranked channel and enhancing the reliability.

Damen et al. ,2003], can still increase the complexity significantly.

**1. Introduction** 

**Retransmission over MIMO** 

Cheng-Ming Chen and Pang-An Ting *Information and Communication Laboratories,* 

**Wireless Systems** 

### **Hybrid ARQ Utilizing Lower Rate Retransmission over MIMO Wireless Systems**

Cheng-Ming Chen and Pang-An Ting *Information and Communication Laboratories, Industrial Technology Research Institute (ITRI), Hsinchu, Taiwan* 

#### **1. Introduction**

Hybrid automatic repeat request (Hybrid ARQ or HARQ), an extension of ARQ that incorporates forward error correction coding (FEC), is a retransmission scheme with errorcontrol method employed in current communications systems. In standard ARQ, redundant bits are added to data to be transmitted using an error-detecting code, e.g., cyclic redundancy check (CRC). The contribution of HARQ is its efficient utilization of the available resources and the provision of reliable services in latest-generation systems.

This chapter focuses on wireless systems using HARQ with emphasis on the multiple-input multiple-output (MIMO) paradigm. In this chapter, the architecture of MIMO transceivers that are based on bit-interleaved coded modulation (BICM) and it employs HARQ is described. MIMO system is an attractive technique that can enhance the spectral efficiency through spatial multiplexing (SM) [Foschini 1995]. However, many wireless environments may suffer from ill-conditioned channels or multipath fadings, which degrade the system performance. The aim of this chapter is to find an efficient MIMO scheme for retransmission to combat the ill-ranked channel and enhancing the reliability.

In IEEE 802.16e [WiMAX 2007] standard, Space Time Coding (STC) subpacket combining, which retransmits with a different MIMO format, has been introduced. One possible way to combine the initial signal and the retransmitted signal is the utilization of Maximum Ratio Combining (MRC) in symbol level based on Virtual Space Time Block Coding (VSTBC) [Gao et. al , 2007]. However, such combining technique works properly only if the channel is quasi-static. In circumstances with high mobility, this technique does not provide satisfactory performance. Another approach is to combine the retransmitted and initial signals using symbol level combining (SLC) before the detector. However, the required buffer in SLC occupies large memory in receivers. Moreover, both SLC receiver architectures are only applicable when the initial and retransmitted symbols are aligned. Thus, they may be impractical for Incremental Redundancy (IR) HARQ [Lin et al. ,1984] and Constellation Rearrangement (CoRe). Although bit level combining (BLC) can solve the problem of huge buffer requirement, using ML-like MIMO detectors, e.g., List Sphere Decoder (LSD) [ Damen et al. ,2003], can still increase the complexity significantly.

Hybrid ARQ Utilizing Lower Rate Retransmission over MIMO Wireless Systems 133

IR or Chase Combining (CC) HARQ [Chase, 1985] as well as the times of reception of NACK

same encoded bits are retransmitted after the request. For IR mode, on the other hand, additional parity check bits are retransmitted. Then, the selected bits, *<sup>s</sup>* **c** , are modulated into

**s** , where *<sup>M</sup> M <sup>2</sup> log* .

*M D M 2k S .S*

*M.S C* **x** .

The signal of ith MIMO encoder output *<sup>i</sup> i ,:* **x x** is STC subpacket encoded based on Table 1, and STC subpacket encoding refers to the VSTBC of previous and current transmitted

It is then followed by a MIMO encoder. The encoder can be either in SM or spatial diversity (S-Div) mode. In SM mode, the MIMO encoder is simply a parser. In S-Div mode, it is Space Time Code (STC) or Space Frequency Code (SFC) encoder. The number of streams of SM and S-Div mode are denoted by *SM* and *SD* , respectively. Note that a stream is defined as each output of the MIMO encoder. The value of *SM* is equivalent to the rate in SM mode. Moreover, in Figure 1, S-Div mode represents rate-1 Alamouti code. Finally, the output of

NACK

QAM Symbols

*<sup>2</sup>* transmission with CC and IR modes. In CC mode, the

NACK counter (Control the MIMO format)

> MIMO encoder

**x**

k/R M.S

QAM Symbols allocated for each stream

M

M

**s**

encoder Bit selection Modulation

NACK counter (Control the selection of bits)

<sup>k</sup> k/R MC k/R k/R

*2k <sup>1</sup>*

*C ,C com <sup>M</sup> plex* 

message.

FEC

Figure 2 gives an example of *<sup>1</sup> <sup>R</sup>*

*M* -ary QAM symbols,

the MIMO encoder is

subpacket, as in 802.16e.

Information bits Mother Code Bits Coded Bits

**b c** *<sup>s</sup>* **c**

AB Y1Y2 W1W2

Fig. 1. MIMO HARQ transmitter diagram of 802.16e

**s** ,

AB AB Y1Y2

To reduce buffer size and power consumption, a lower rate MIMO mode in retransmission request, termed as Lower Rate Retransmission (LRR) Scheme [Chen et. al, 2009], is proposed. In this chapter, we define the rate as how much information can be transmitted in single time-frequency resource unit. Two examples of LRR schemes are introduced. For the first scheme, it applies rate-2 SM for initial transmission, and rate-1 STBC [Alamouti ,1998] or Space Frequency Block Coding (SFBC) [Kaiser ,2003] is used for retransmission. For the second scheme, rate-3 or rate-4 SM are leveraged for initial transmission, and a lower rate SM scheme, i.e., rate-2 or rate-3, is employed for retransmission. In order not to decrease the spectral efficiency, only partial coded bits are retransmitted in the proposed schemes.

In LRR, with fewer transmit antennae in retransmission, the acquired transmit power gain could be used to retransmit higher modulation symbols and keep the total retransmitted bits as close to initial transmission as possible. We tabulated this scheme as Lower Rate Retransmission combine with modulation step up (LRRMSU).

The notations of this paper are explained as following: Superscripts *T* and *H* indicate matrix transpose and hermitian, respectively. Superscripts *\** indicates complex conjugate operation. Uppercase boldface denotes a matrix while lowercase boldface denotes a vector. **I***N* denotes the *N N* Identity matrix. *i ,j A* and *:,j A* represent the element of ith row and jth column of matrix **A** , and jth column of matrix A, respectively. A circularly symmetric complex Gaussian vector **a** with mean **m** and covariance matrix **R** is denoted as **a** ∼ NC [ **m** , **R** ]. Finally, *nTX* , *Mt* and *Mr* refer to the n-th transmission, number of transmit antenna and receive antenna, respectively.

This article is organized as follows. In the next section the architecture of a single-input single-output (SISO) transceiver using BICM and HARQ is presented. The following section narrates conventional MIMO HARQ schemes. The LRR schemes are then elaborated with simpler receiver implementations. The following section contains some discussion of MIMO system design based on the employed HARQ scheme, receiver complexity, and storage requirements. To keep the number of retransmission bits, near to that of the initial transmission a LRRMSU scheme is illustrated to enhance the system throughput. Finally, some concluding remarks are provided.

#### **2. MIMO HARQ scheme in 802.16**

Figure 1 shows the block diagram of MIMO HARQ in 802.16e transmitter. It is illustrated that *k* information bits **b** *, k 1 b |b 0,1* **b** are fit into one forward error control coding (FEC) block. These information bits are regarded as systematic bits *AB* . The mother code rate *RMC* of the Convolutional Turbo Code (CTC) in 802.16e is *<sup>1</sup> <sup>3</sup>* . After CTC encoding, the encoded bit length is multiplied by 3, and the output *3k 1 , b |b 0,1* **c c** , is consists of systematic bits *AB* , parity bits *Y Y1 2* and *W W1 2* . In this example, the code rate *R* is *<sup>1</sup> 2* , hence in the initial transmission, *W W1 2* are punctured in the bit selection block, and the remained encoded bits are *2k 1 s s , b |b 0,1* **c c** . The bit selection procedure depends on

To reduce buffer size and power consumption, a lower rate MIMO mode in retransmission request, termed as Lower Rate Retransmission (LRR) Scheme [Chen et. al, 2009], is proposed. In this chapter, we define the rate as how much information can be transmitted in single time-frequency resource unit. Two examples of LRR schemes are introduced. For the first scheme, it applies rate-2 SM for initial transmission, and rate-1 STBC [Alamouti ,1998] or Space Frequency Block Coding (SFBC) [Kaiser ,2003] is used for retransmission. For the second scheme, rate-3 or rate-4 SM are leveraged for initial transmission, and a lower rate SM scheme, i.e., rate-2 or rate-3, is employed for retransmission. In order not to decrease the spectral efficiency, only partial coded bits are retransmitted in the proposed schemes.

In LRR, with fewer transmit antennae in retransmission, the acquired transmit power gain could be used to retransmit higher modulation symbols and keep the total retransmitted bits as close to initial transmission as possible. We tabulated this scheme as Lower Rate

The notations of this paper are explained as following: Superscripts *T* and *H* indicate matrix transpose and hermitian, respectively. Superscripts *\** indicates complex conjugate operation. Uppercase boldface denotes a matrix while lowercase boldface denotes a vector. **I***N* denotes the *N N* Identity matrix. *i ,j A* and *:,j A* represent the element of ith row and jth column of matrix **A** , and jth column of matrix A, respectively. A circularly symmetric complex Gaussian vector **a** with mean **m** and covariance matrix **R** is denoted as **a** ∼ NC [ **m** , **R** ]. Finally, *nTX* , *Mt* and *Mr* refer to the n-th transmission, number of

This article is organized as follows. In the next section the architecture of a single-input single-output (SISO) transceiver using BICM and HARQ is presented. The following section narrates conventional MIMO HARQ schemes. The LRR schemes are then elaborated with simpler receiver implementations. The following section contains some discussion of MIMO system design based on the employed HARQ scheme, receiver complexity, and storage requirements. To keep the number of retransmission bits, near to that of the initial transmission a LRRMSU scheme is illustrated to enhance the system throughput. Finally,

Figure 1 shows the block diagram of MIMO HARQ in 802.16e transmitter. It is illustrated that *k* information bits **b** *, k 1 b |b 0,1* **b** are fit into one forward error control coding (FEC) block. These information bits are regarded as systematic bits *AB* . The mother code

encoded bit length is multiplied by 3, and the output *3k 1 , b |b 0,1* **c c** , is consists of

hence in the initial transmission, *W W1 2* are punctured in the bit selection block, and the

*s s , b |b 0,1* **c c** . The bit selection procedure depends on

systematic bits *AB* , parity bits *Y Y1 2* and *W W1 2* . In this example, the code rate *R* is *<sup>1</sup>*

*<sup>3</sup>* . After CTC encoding, the

*2* ,

rate *RMC* of the Convolutional Turbo Code (CTC) in 802.16e is *<sup>1</sup>*

remained encoded bits are *2k 1*

Retransmission combine with modulation step up (LRRMSU).

transmit antenna and receive antenna, respectively.

some concluding remarks are provided.

**2. MIMO HARQ scheme in 802.16** 

IR or Chase Combining (CC) HARQ [Chase, 1985] as well as the times of reception of NACK message.

Fig. 1. MIMO HARQ transmitter diagram of 802.16e

Figure 2 gives an example of *<sup>1</sup> <sup>R</sup> <sup>2</sup>* transmission with CC and IR modes. In CC mode, the same encoded bits are retransmitted after the request. For IR mode, on the other hand, additional parity check bits are retransmitted. Then, the selected bits, *<sup>s</sup>* **c** , are modulated into *M* -ary QAM symbols,

$$\mathbf{s} \text{ \{ \mathbf{s} \in \overset{2k}{\mathbf{C}^M}, \mathbf{C} \in \text{complex} \} \text{ \textquotedblleft where } M = \log\_2^{\overline{M}} \text{ \textquotedblright}$$

It is then followed by a MIMO encoder. The encoder can be either in SM or spatial diversity (S-Div) mode. In SM mode, the MIMO encoder is simply a parser. In S-Div mode, it is Space Time Code (STC) or Space Frequency Code (SFC) encoder. The number of streams of SM and S-Div mode are denoted by *SM* and *SD* , respectively. Note that a stream is defined as each output of the MIMO encoder. The value of *SM* is equivalent to the rate in SM mode. Moreover, in Figure 1, S-Div mode represents rate-1 Alamouti code. Finally, the output of the MIMO encoder is

$$\mathbf{x} \in \overset{2k}{\mathbf{C}^{M.S\_M}} \ast \mathbf{S}\_M.\mathbf{S}\_D$$

The signal of ith MIMO encoder output *<sup>i</sup> i ,:* **x x** is STC subpacket encoded based on Table 1, and STC subpacket encoding refers to the VSTBC of previous and current transmitted subpacket, as in 802.16e.

Hybrid ARQ Utilizing Lower Rate Retransmission over MIMO Wireless Systems 135

in Table 2. Since an open loop system is being considered, thus the transmitter does not possess channel state information (CSI). Note that LRR can be leveraged in association with

*Mt 2* rate-2 SM rate-1 SFBC/STBC

In order to maintain the same spectral efficiency as the conventional HARQ schemes, fewer bits are encoded by LRR MIMO encoder. Although the number of retransmitted bit is reduced, the reliability is improved. Since only a portion of bits is retransmitted, the bit selection should be modified. An example of retransmissions with coding rate *<sup>1</sup> <sup>R</sup>*

CC mode IR mode

*<sup>2</sup>* in CC

W1W2

AB Y1Y2

AB

Circular Buffer Size

Y1Y2

*Mt 3* rate-3 SM rate-2 SM *Mt 4* rate-4 SM rate-3 SM

a stream-to-antenna mapping technique such as precoding or antenna selection.

Antenna Initial TX=1TX Request TX

AB Y1Y2

AB

Circular Buffer Size

Y1Y2

In this section, four types of MIMO receivers are elaborated. Firstly, we illustrate the combining methods for both VSTBC and STBC/SFBC in symbol and bit levels. Then, two SM detection algorithms in BLC are described: Soft Linear Minimum Mean Square Error (LMMSE)[ Lee & Sundberg, 2007] algorithm and LSD algorithm. All algorithms use soft decision information generated from CSI. Finally, a complexity analysis is carried out in a

Fig. 3. Different Bit selection methods for CC (left) and IR (right) with LRR scheme.

reTX

Table 2. MIMO mode selection in LRR

and IR modes are shown as Figure 3.

1TX

2TX

3TX

4TX AB

SM

STBC/SFBC

STBC/SFBC

STBC/SFBC

**4. Receiver architectures** 

rate-2 MIMO HARQ scheme.

Fig. 2. Different Bit selection manners for CC (left) and IR (right)


Table 1. STC subpacket encoding

#### **3. Lower Rate Retransmission (LRR) scheme**

In practice, due to the propagation mechanisms, MIMO system may be suffered from high spatial correlation, which degrades the system capacity. If there is no feedback information regarding the channel rank, it is always a good approach to retransmit with a lower rate MIMO mode to provide a robust transmission. For instance, a rate-1 SFBC or STBC is recommended for retransmission for initial transmission in rate-2 SM mode. For rate-3 and rate-4 initial transmission in SM mode, on the other hand, rate-2 and rate-3 SM mode are recommended for retransmission. A list of possible MIMO mode selection for LRR is shown in Table 2. Since an open loop system is being considered, thus the transmitter does not possess channel state information (CSI). Note that LRR can be leveraged in association with a stream-to-antenna mapping technique such as precoding or antenna selection.


Table 2. MIMO mode selection in LRR

134 Advanced Transmission Techniques in WiMAX

CC mode IR mode

W1W2

*1*

*2 s s* 

> *1 2*

 

*s s s s*

*3 4*

*s s s*

 

*i*

**x**

**x**

*i*

**x**

*\* 2 i \* 1*

*\* 2 \* i 1 \* 3*

*\* 2 \* 1 i \* 4 \* 3*

*s s s s*

*s s s* 

*s s* 

**x**

**x**

**x**

Y1Y2 W1W2

AB Y1Y2

AB

Circular Buffer Size

AB Y1Y2

1TX

2TX

3TX

reTX

*Mt 3*

*Mt 4*

Table 1. STC subpacket encoding

*Mt <sup>2</sup> <sup>1</sup>*

AB Y1Y2

AB Y1Y2

Circular Buffer Size

Fig. 2. Different Bit selection manners for CC (left) and IR (right)

*i*

*i*

**3. Lower Rate Retransmission (LRR) scheme** 

*2 s s* **x**

> > *1 2*

*s s s s*

 **x**

*3 4*

In practice, due to the propagation mechanisms, MIMO system may be suffered from high spatial correlation, which degrades the system capacity. If there is no feedback information regarding the channel rank, it is always a good approach to retransmit with a lower rate MIMO mode to provide a robust transmission. For instance, a rate-1 SFBC or STBC is recommended for retransmission for initial transmission in rate-2 SM mode. For rate-3 and rate-4 initial transmission in SM mode, on the other hand, rate-2 and rate-3 SM mode are recommended for retransmission. A list of possible MIMO mode selection for LRR is shown

*s s s* **x**

4TX AB Y1Y2 AB Y1Y2

Antenna Initial TX=1TX Even TX Odd TX

In order to maintain the same spectral efficiency as the conventional HARQ schemes, fewer bits are encoded by LRR MIMO encoder. Although the number of retransmitted bit is reduced, the reliability is improved. Since only a portion of bits is retransmitted, the bit selection should be modified. An example of retransmissions with coding rate *<sup>1</sup> <sup>R</sup> <sup>2</sup>* in CC and IR modes are shown as Figure 3.

Fig. 3. Different Bit selection methods for CC (left) and IR (right) with LRR scheme.

#### **4. Receiver architectures**

In this section, four types of MIMO receivers are elaborated. Firstly, we illustrate the combining methods for both VSTBC and STBC/SFBC in symbol and bit levels. Then, two SM detection algorithms in BLC are described: Soft Linear Minimum Mean Square Error (LMMSE)[ Lee & Sundberg, 2007] algorithm and LSD algorithm. All algorithms use soft decision information generated from CSI. Finally, a complexity analysis is carried out in a rate-2 MIMO HARQ scheme.

Hybrid ARQ Utilizing Lower Rate Retransmission over MIMO Wireless Systems 137

*0 0 H odd ^ ^ \* even 1 1*

(7)

*s1* , respectively. The

(9)

(10)

*s0* and *^*

**H H** (8)

 **<sup>y</sup> <sup>H</sup> y**

 *^ 0 H*

In equation 7, if the channel is not static for each retransmission, the orthogonality of Alamouti Code is destroyed and the interference is thereby induced. The mathematical

*h hhhh hhh*

*odd\* odd odd\* odd even even\* even 00 0 10 1 01 0 11*

*<sup>2</sup> 2 2*

*<sup>2</sup> odd\* odd Eh h ij mn* ,

*<sup>2</sup> \* E h h ,ij mn ij mn* .

The overall receiver architecture is shown in Figure 4. The scheme requires not only symbol level buffer (SLB), but also bit level buffer (BLB), where the BLB is used to store loglikelihood ratios (LLRs). For example, in 3TX, retransmitted packet cannot be combined by MRC with previous symbol values using VSTBC format, hence only previous LLRs is required to be stored in BLB in this retransmission. The SLB and BLB cannot be shared with

*1 2 4N* 

*<sup>2</sup> odd\* Eh h ij mn* and

*<sup>8</sup> SNR*

*o*

*hn hn hn h*

*odd\* odd \* odd\* odd \* 00 01 01 00 10 11 11 10*

*even\* n , <sup>1</sup>*

where the second term is the interference induced from signal *<sup>1</sup>* **s** . The equivalent SNR of

*^ ^ 22 2 2 odd odd even even 0 0 00 10 01 11 0*

*hs h h h h s*

*diag*

*h1* can also be obtained by MRC in eq. 8.

*^ ^*

*h s*

*h s*

*^ 1*

*h*

*1 hh hh*

*00 01 10 11 1*

*hh hh s*

*2 odd\* odd odd\* odd 00 01 10 11*

*\* \**

each other, because both of them are required for 4TX combining.

*h*

 

*h0* and *^*

The MRC for channel operates as following:

with the assumptions of cross correlation

*h s 1 1* are the soft detection symbols of *^*

Where *^ ^*

*h s 0 0* and *^ ^*

equivalent channel gain *^*

description is as follows:

signal *<sup>0</sup>* **s** is:

#### **4.1 VSTBC-MRC with SLC and STBC/SFBC-MRC with BLC**

A flat-fading MIMO system can be expressed as:

$$\mathbf{y} = \underline{\mathbf{H}} \mathbf{P} \mathbf{x} + \mathbf{n} = \mathbf{H} \mathbf{x} + \mathbf{n},\tag{1}$$

where **H** is the *Mr t M* MIMO channel matrix of an OFDM subcarrier with unitary power complex Gaussian elements, **x** is the *S 1 <sup>M</sup>* transmitted signal vector with unit total transmission power per subcarrier, and **y** is the *Mr 1* received signal vector. Moreover, **<sup>P</sup>** is the *Mt M <sup>S</sup>* precoding matrix, and *C oMr ~ N ,N .* **n 0I** is the *Mr <sup>1</sup>* noise vector with complex Gaussian elements. Without loss of generality, a MIMO scenario with *Mt r M 2* and identity matrix **P** is described.

The signal in the odd transmission (as defined in Table 1) is:

$$\mathbf{x}\_{\alpha dd} = \begin{bmatrix} \mathbf{x}\_0 \\ \mathbf{x}\_1 \end{bmatrix} \tag{2}$$

And the signal in the even transmission is:

$$\mathbf{x}\_{even} = \begin{bmatrix} -\boldsymbol{\alpha}\_1^\* \\ \boldsymbol{\alpha}\_0^\* \end{bmatrix} \tag{3}$$

Here, subscripts *odd* and *even* indicate different subpacket information at odd and even retransmission. Moreover, the initial transmission is started with the first odd transmission. Hence, a pair of received signal in subpackets can be formed as:

$$\mathbf{y}\_{\alpha dd} = \mathbf{H}\_{\alpha dd} \mathbf{x}\_{\alpha dd} + \mathbf{n}\_{\alpha dd} \tag{4}$$

$$\mathbf{y}\_{\text{even}} = \left(\mathbf{p}\mathbf{H}\_{\text{odd}} + \Delta\mathbf{H}\right)\mathbf{x}\_{\text{even}} + \mathbf{n}\_{\text{even}} \tag{5}$$

Here, we assume **H HH** *even odd* and *<sup>r</sup> <sup>2</sup> ~N , 1 C M* **H0 I** with *0 1* is induced from Doppler effect. Equation 4 and 5 can be further stacked as:

$$
\begin{bmatrix}
\mathbf{y}\_{odd} \\
\mathbf{y}\_{even}^\*
\end{bmatrix} = \overline{\mathbf{H}} \mathbf{x}\_{odd} + \begin{bmatrix}
\mathbf{n}\_{odd} \\
\mathbf{n}\_{even}^\*
\end{bmatrix} \tag{6}
$$

Here,

$$
\overline{\mathbf{H}} = \begin{bmatrix}
\left[\mathbf{H}\_{\text{odd}}\right]\_{;,1} & \left[\mathbf{H}\_{\text{odd}}\right]\_{;,2} \\
\left[\mathbf{H}\_{\text{even}}\right]\_{;,2}^\* & -\left[\mathbf{H}\_{\text{even}}\right]\_{;,1}^\*
\end{bmatrix}'
$$

and the combined signal vector by MRC is shown in eq. 7.

The MRC for signal operates as following:

$$
\begin{bmatrix}
\stackrel{\frown}{h\_{IO}}\stackrel{\frown}{s\_{O}}\\\
\stackrel{\frown}{h\_{I}s\_{I}}
\end{bmatrix} = \overline{\mathbf{H}}^{H} \begin{bmatrix}
\mathbf{y}\_{cold} \\
\mathbf{y}\_{evm}
\end{bmatrix} \tag{7}
$$

Where *^ ^ h s 0 0* and *^ ^ h s 1 1* are the soft detection symbols of *^ s0* and *^ s1* , respectively. The equivalent channel gain *^ h0* and *^ h1* can also be obtained by MRC in eq. 8.

The MRC for channel operates as following:

136 Advanced Transmission Techniques in WiMAX

where **H** is the *Mr t M* MIMO channel matrix of an OFDM subcarrier with unitary power complex Gaussian elements, **x** is the *S 1 <sup>M</sup>* transmitted signal vector with unit total transmission power per subcarrier, and **y** is the *Mr 1* received signal vector. Moreover, **<sup>P</sup>** is the *Mt M <sup>S</sup>* precoding matrix, and *C oMr ~ N ,N .* **n 0I** is the *Mr <sup>1</sup>* noise vector with complex Gaussian elements. Without loss of generality, a MIMO scenario with *Mt r M 2*

*0*

*1 x x* 

> *\* 1*

*x x* 

*odd*

*even \* 0*

Here, subscripts *odd* and *even* indicate different subpacket information at odd and even retransmission. Moreover, the initial transmission is started with the first odd transmission.

> *odd odd \* \* odd even even* **y n Hx y n**

 **H H**

**H H** ,

 *odd odd :,1 :,2 \* \* even :,2 even :,1*

**y HPx n Hx n** *,* (1)

**x** (2)

**x** (3)

**y Hx n** *odd odd odd odd* (4)

*<sup>2</sup> ~N , 1 C M* **H0 I** with *0 1* is induced

(6)

**y H Hx n** *even odd even even* (5)

**4.1 VSTBC-MRC with SLC and STBC/SFBC-MRC with BLC** 

The signal in the odd transmission (as defined in Table 1) is:

Hence, a pair of received signal in subpackets can be formed as:

Here, we assume **H HH** *even odd* and *<sup>r</sup>*

from Doppler effect. Equation 4 and 5 can be further stacked as:

**H**

and the combined signal vector by MRC is shown in eq. 7.

The MRC for signal operates as following:

A flat-fading MIMO system can be expressed as:

and identity matrix **P** is described.

And the signal in the even transmission is:

Here,

$$
\begin{bmatrix}
\hat{\boldsymbol{h}}\_{\boldsymbol{h}\_{0}} \\
\hat{\boldsymbol{h}}\_{\boldsymbol{h}\_{1}}
\end{bmatrix} = \operatorname\*{diag}\left(\overline{\mathbf{H}}^{H}\,\overline{\mathbf{H}}\right) \tag{8}
$$

In equation 7, if the channel is not static for each retransmission, the orthogonality of Alamouti Code is destroyed and the interference is thereby induced. The mathematical description is as follows:

$$\begin{aligned} \stackrel{\circ}{16}\stackrel{\circ}{16} &= \left( \left| h\_{00}^{\alpha d d} \right|^2 + \left| h\_{10}^{\alpha d d} \right|^2 + \left| h\_{01}^{\epsilon even} \right|^2 + \left| h\_{11}^{\epsilon even} \right|^2 \right) \mathbf{s}\_0 + \\ & \left( \left( 1 - \rho^2 \right) \left( h\_{00}^{\alpha d d^\*} h\_{01}^{\alpha d d} + h\_{10}^{\alpha d d^\*} h\_{11}^{\alpha d d} \right) - \\ & \left( \rho \left( h\_{00}^{\alpha d d^\*} \Delta h\_{01} + h\_{01}^{\alpha d d} \Delta h\_{00}^{\dagger} + h\_{10}^{\alpha d d^\*} \Delta h\_{11} + h\_{11}^{\alpha d d} \Delta h\_{10}^{\dagger} \right) - \\ & \left( \Delta h\_{00}^{\star} \Delta h\_{01} + \Delta h\_{10}^{\star} \Delta h\_{11} \right) \right) \mathbf{s}\_1 + \\ & \left( h\_{00}^{\alpha d d^\*} n\_0^{\alpha d d} + h\_{10}^{\alpha d d^\*} n\_1^{\alpha d d} + h\_{01}^{\alpha c u} n\_0^{\alpha c e n^\*} + h\_{11}^{\alpha e u} n\_1^{\alpha e n^\*} \right), \end{aligned} \tag{9}$$

where the second term is the interference induced from signal *<sup>1</sup>* **s** . The equivalent SNR of signal *<sup>0</sup>* **s** is:

$$SNR = \frac{8}{\left(\left(1-\rho^2\right)^2 \gamma + 2\rho^2 \mathfrak{B} + \alpha\right) + 4N\_o} \tag{10}$$

with the assumptions of cross correlation

$$E\left\{\left|h\_{ij}^{\alpha dd^\*}h\_{mn}^{\alpha dd}\right|^2\right\}=\gamma\;E\left\{\left|h\_{ij}^{\alpha dd^\*}\Delta h\_{mn}\right|^2\right\}=\emptyset \quad \text{and} \quad E\left\{\left|\Delta h\_{ij}^{\alpha ddd^\*}\Delta h\_{mn}\right|^2\right\}=\alpha\text{,}\;ij\neq mn\;.$$

$$E\left\{\left|\Delta h\_{ij}^{\alpha}\Delta h\_{mn}\right|^2\right\}=\alpha\text{,}\;ij\neq mn\;.$$

The overall receiver architecture is shown in Figure 4. The scheme requires not only symbol level buffer (SLB), but also bit level buffer (BLB), where the BLB is used to store loglikelihood ratios (LLRs). For example, in 3TX, retransmitted packet cannot be combined by MRC with previous symbol values using VSTBC format, hence only previous LLRs is required to be stored in BLB in this retransmission. The SLB and BLB cannot be shared with each other, because both of them are required for 4TX combining.

Hybrid ARQ Utilizing Lower Rate Retransmission over MIMO Wireless Systems 139

In BLC, there is no restrictions on symbol-alignment for the signals of odd and even retransmissions. Based on the system model in equation 1, and signal vector **x** here is the same as **x***odd* , the detection signal can be expressed as in 13. The LMMSE for signal operates

<sup>H</sup> *<sup>1</sup> <sup>H</sup> S NM o*

The equivalent channel gain is the inverse of diagonal terms of MSE matrix. The LMMSE for

 *^ <sup>1</sup> <sup>0</sup> <sup>H</sup>*

*<sup>h</sup> diag SN , <sup>1</sup>*

Received Pattern Bit level Buffer

There are many simplified LLR algorithms for soft values computation for a SM system in the existing literature. Here, we focus on an exhaustive search which results in no penalty in performance. The LLR of the *k*th bit on the transmitted symbol vector **x** (which

> *<sup>k</sup> k ,1 k ,0 kk1 kk1 t o <sup>1</sup> LLR min min , M N*

*2 2*

**y Hx y Hx** (15)

*^ 1*

*h*

The LMMSE channel gain process

SM detection

*1*

*M o*

*h1* are the equivalent soft CSI gains and the overall scheme is shown in

Soft Bit demapping

> 0 1 *s s*

Soft Demapper

 0 1 *h h* 

**x HH I Hy** (13)

**HH I** (14)

**LLR**

FEC Decoder

Former LLRs

+

Outer

**4.2 Soft LMMSE with BLC** 

channel gain operates as:

*h0* and *^*

**y**

**H**

Fig. 6. Soft LMMSE processing with BLC

**4.3 LSD with BLC** 

contains *S M <sup>M</sup>* bits) is:

as:

where *^*

Figure 6.

Fig. 4. Subpacket Combining Process

For the proposed LRR scheme with rate-2 SM initial transmission, the scheme is simply tantamount to a SFBC after rate reduction, and it can be detected by MRC method and then combined with previous LLRs in BLC. The SFBC formats of the retransmitted subpacket are expressed as:

$$\mathbf{x}\_{sub\_0} = \begin{bmatrix} \mathbf{x}\_0 \\ \mathbf{x}\_1 \end{bmatrix} \tag{11}$$

and

$$\mathbf{x}\_{sub\_l} = \begin{bmatrix} -\boldsymbol{\chi}\_1^\* \\ \boldsymbol{\chi}\_0^\* \end{bmatrix} \tag{12}$$

where subscripts *<sup>0</sup> sub* and *<sup>1</sup> sub* indicate the subcarrier indices 0 and 1, respectively. The detection is performed for every two consecutive subcarriers and the former detected LLRs is stored in BLB, hence SLB is not required. The overall receiver architecture is shown in Figure 5.

Fig. 5. SFBC MRC with BLC

#### **4.2 Soft LMMSE with BLC**

138 Advanced Transmission Techniques in WiMAX

 0 0 1 1 *h s h s* 

 0 1 *h h* 

For the proposed LRR scheme with rate-2 SM initial transmission, the scheme is simply tantamount to a SFBC after rate reduction, and it can be detected by MRC method and then combined with previous LLRs in BLC. The SFBC formats of the retransmitted subpacket are

*0*

*1*

The MRC channel gain process

MRC

*sub \* 0*

where subscripts *<sup>0</sup> sub* and *<sup>1</sup> sub* indicate the subcarrier indices 0 and 1, respectively. The detection is performed for every two consecutive subcarriers and the former detected LLRs is stored in BLB, hence SLB is not required. The overall receiver architecture is shown in

> 0 0 1 1 *h s h s*

> > 0 1 *h h*

*sub*

*0*

*1 x x* 

> *\* 1*

Soft Demapper

Soft Bit demapping

*x x* 

Soft Demapper

**LLR**

**x** (11)

**x** (12)

**LLR**

FEC Decoder

Bit level Buffer

Former LLRs

+

Outer

FEC Decoder

Bit level Buffer

Former LLRs

+

Outer

Soft Bit demapping

The MRC channel gain process

Subpacket Combining

*even*

Received Pattern

Symbol level Buffer

*odd*

**y**

**H**

*odd*

**y**

**H**

expressed as:

and

Figure 5.

*even*

Fig. 4. Subpacket Combining Process

0 0

1 1

*sub sub*

**y H**

Fig. 5. SFBC MRC with BLC

*sub sub*

Received Pattern

**y H** In BLC, there is no restrictions on symbol-alignment for the signals of odd and even retransmissions. Based on the system model in equation 1, and signal vector **x** here is the same as **x***odd* , the detection signal can be expressed as in 13. The LMMSE for signal operates as:

$$\hat{\mathbf{x}} = \left(\mathbf{H}^H \mathbf{H} + \mathbf{S}\_M N\_\sigma \mathbf{I}\right)^{-1} \mathbf{H}^H \mathbf{y} \tag{13}$$

The equivalent channel gain is the inverse of diagonal terms of MSE matrix. The LMMSE for channel gain operates as:

$$\begin{bmatrix} \hat{I} \\ \hat{h}\_{0} \\ \frac{1}{\hat{h}\_{1}} \end{bmatrix} = \text{diag}\left(\left(\mathbf{H}^{H}\mathbf{H} + \mathbf{S}\_{M}\mathbf{N}\_{o}\mathbf{I}\right)^{-1}\right) \tag{14}$$

where *^ h0* and *^ h1* are the equivalent soft CSI gains and the overall scheme is shown in Figure 6.

Fig. 6. Soft LMMSE processing with BLC

#### **4.3 LSD with BLC**

There are many simplified LLR algorithms for soft values computation for a SM system in the existing literature. Here, we focus on an exhaustive search which results in no penalty in performance. The LLR of the *k*th bit on the transmitted symbol vector **x** (which contains *S M <sup>M</sup>* bits) is:

$$LLRR\_k \approx \frac{1}{M\_t N\_o} \times \left( -\min\_{k \le k+1} \left\| \mathbf{y} - \mathbf{H} \mathbf{x}\_{k,1} \right\|^2 + \min\_{k \le k-1} \left\| \mathbf{y} - \mathbf{H} \mathbf{x}\_{k,0} \right\|^2 \right), \tag{15}$$

Hybrid ARQ Utilizing Lower Rate Retransmission over MIMO Wireless Systems 141

**HH I Hy** , needs further four CMLs. Hence, there are 1344 CMLs in total.

**5. Lower Rate Retransmission (LRR) combined with Modulation Step Up** 

To recapitulate, our proposed scheme reduces the complexity by about 50% and 70% as

There are several pros of the LRR scheme such as 1. a robust retransmission, because the inter-stream interference is reduced, 2. the transmitter side acquires additional transmit power gain for each retransmitted stream, 3. frequency and spatial diversity is gained, because different resource allocation will be automatically guaranteed. However, one possible cons, the total retransmitted bits will be reduced as compare to number of bits in initial transmission, might degrade its performance in high coding rate scenarios. To overcome this deficiency, higher order modulation or called modulation step up is introduced and combined with the LRR scheme in Figure 8. In the initial transmission, the transmission mode operates with 4 transmission antennae with QPSK in each stream. In retransmission, the number of transmission antennae is reduced to 3 but the modulation order is step up to 16 QAM. Therefore, we keep the number of retransmission bits very close

**HH I** with *<sup>H</sup>* **H** takes eight CMLs, and finally,

compared to VSTBC-MRC with SLC and soft LMMSE with BLC, respectively.

multiplication of *<sup>1</sup> <sup>H</sup> S NM o*

*<sup>1</sup> H H S NM o*

**(LRRMSU) scheme** 

to that of traditional scheme.

Fig. 8. LRRMSU retransmission scheme

where *k ,1* **x** and *k ,0* **x** are the symbol vectors with *k*th bit equals+1 and −1, respectively. The block diagram of LSD with BLC is shown in Figure 7:

Fig. 7. LSD processing with BLC

#### **4.4 Complexity analysis**

An example based on 802.16e Partial Usage of Subchannels (PUSC) format is given here to analyze the complexity of different schemes under our consideration. The resource unit is constructed with one slot with 48 data subcarriers. For sake of simplicity, only the number of complex multiplication operations (CMLs) is considered. A real division operation is assumed to be equivalent to one CML. Besides, the complexity of LSD with BLC is not examined, because its complexity depends on the modulation order and it is undoubtable to be more complicated than the other detectors. The complexity of a MIMO receiver with *Mt r M 2* is shown as follows.

1. VSTBC-MRC with SLC:

In this case, there are 48 subcarriers in total to be implemented with the MRC operation, and each of them takes eight CMLs in MRC for signal and eight CMLs (only diagonal elements are needed) in MRC for channel. Hence, it takes 768 CMLs in total.

2. STBC/SFBC-MRC with BLC:

Every 2-subcarrier pair is to be implemented by one MRC operation in this case, hence the number of required operations is half of VSTBC-MRC with SLC. Thus, the total number of required operations is 384 CMLs.

3. Soft LMMSE with BLC:

In this scenario, we should consider 48 subcarriers with *2 2* matrix operations in MMSE filtering. Each subcarrier requires eight CMLs to compute *<sup>H</sup>* **H H** , eight CMLs for matrix inversion (two CMLs for the determinant and six CMLs for the real divisions), the

SM detection

Received Pattern Bit level Buffer

The LSD process

An example based on 802.16e Partial Usage of Subchannels (PUSC) format is given here to analyze the complexity of different schemes under our consideration. The resource unit is constructed with one slot with 48 data subcarriers. For sake of simplicity, only the number of complex multiplication operations (CMLs) is considered. A real division operation is assumed to be equivalent to one CML. Besides, the complexity of LSD with BLC is not examined, because its complexity depends on the modulation order and it is undoubtable to be more complicated than the other detectors. The complexity of a MIMO receiver with

In this case, there are 48 subcarriers in total to be implemented with the MRC operation, and each of them takes eight CMLs in MRC for signal and eight CMLs (only diagonal elements

Every 2-subcarrier pair is to be implemented by one MRC operation in this case, hence the number of required operations is half of VSTBC-MRC with SLC. Thus, the total number of

In this scenario, we should consider 48 subcarriers with *2 2* matrix operations in MMSE filtering. Each subcarrier requires eight CMLs to compute *<sup>H</sup>* **H H** , eight CMLs for matrix inversion (two CMLs for the determinant and six CMLs for the real divisions), the

are needed) in MRC for channel. Hence, it takes 768 CMLs in total.

−

FEC Decoder

Former LLRs

+

**LLR**

Outer

1, respectively. The

where *k ,1* **x** and *k ,0* **x** are the symbol vectors with *k*th bit equals+1 and

block diagram of LSD with BLC is shown in Figure 7:

**y**

**H**

Fig. 7. LSD processing with BLC

*Mt r M 2* is shown as follows.

2. STBC/SFBC-MRC with BLC:

required operations is 384 CMLs.

3. Soft LMMSE with BLC:

1. VSTBC-MRC with SLC:

**4.4 Complexity analysis** 

multiplication of *<sup>1</sup> <sup>H</sup> S NM o* **HH I** with *<sup>H</sup>* **H** takes eight CMLs, and finally, *<sup>1</sup> H H S NM o* **HH I Hy** , needs further four CMLs. Hence, there are 1344 CMLs in total.

To recapitulate, our proposed scheme reduces the complexity by about 50% and 70% as compared to VSTBC-MRC with SLC and soft LMMSE with BLC, respectively.

#### **5. Lower Rate Retransmission (LRR) combined with Modulation Step Up (LRRMSU) scheme**

There are several pros of the LRR scheme such as 1. a robust retransmission, because the inter-stream interference is reduced, 2. the transmitter side acquires additional transmit power gain for each retransmitted stream, 3. frequency and spatial diversity is gained, because different resource allocation will be automatically guaranteed. However, one possible cons, the total retransmitted bits will be reduced as compare to number of bits in initial transmission, might degrade its performance in high coding rate scenarios. To overcome this deficiency, higher order modulation or called modulation step up is introduced and combined with the LRR scheme in Figure 8. In the initial transmission, the transmission mode operates with 4 transmission antennae with QPSK in each stream. In retransmission, the number of transmission antennae is reduced to 3 but the modulation order is step up to 16 QAM. Therefore, we keep the number of retransmission bits very close to that of traditional scheme.

Fig. 8. LRRMSU retransmission scheme

Hybrid ARQ Utilizing Lower Rate Retransmission over MIMO Wireless Systems 143

Fig. 9. VSTBC SLC v.s. SFBC BLC in 2TX PB3 CC

Fig. 10. VSTBC SLC v.s. SFBC BLC in 2TX VA60 CC

#### **6. Numerical results**

In order to verify the superiority of the proposed scheme, the simulation based on a low correlation MIMO model [WiMAX ,2007] with *Mt r M 2* is undertaken here. In particular, we show two examples of comparison in this paper: *VSTBC in SLC* v.s. *SFBC in BLC* and *SFBC in BLC* v.s. *SM in BLC*. The delay profile of each path is evaluated under ITU-R [ITU-R ,2000] Pedestrian Type-B 3km/hr (PB3) or Vehicular Type-A 60km/hr (VA60). Furthermore, PUSC with 10Mhz bandwidth is assumed and the coding scheme is based on 802.16e CTC. In addition, we postulate that the receiver has perfect CSI. The HARQ round trip interval is 10ms, and the subpacket will be shifted by 3 subchannel length to gain higher diversity in frequency domain. The subpacket size for each coding rate is summarized in Table 3, where *NEP* is the number of information bits before feeding into FEC encoder. Note that we concentrate on CC mode and the comparison with IR mode is beyond our scope due to page restrictions.


Table 3. *NEP* size of different code rate with CC

#### **Example 1: VSTBC in SLC v.s. SFBC in BLC.**

We focus on simulation results of PB3 and VA60. In 1TX, the receiver is a soft MMSE detector for both cases, and in 2TX the VSTBC-MRC or SFBC-MRC is used. Simulation results of 2TX packet error rate (PER) in Figure 9 and 10 show that our proposed scheme has poorer error performance in low mobility. However, the interference terms in equation 9 will impact the performance of VSTBC in SLC scheme as the mobility increases. Hence, the proposed scheme becomes superior in moderate and higher speed scenarios. Nevertheless, a Doppler estimator is generally not available at the receiver; hence conventional 802.16e cannot be guaranteed to be superior at all mobility levels. The PER in 3TX is not evaluated, because the retransmitted subpacket cannot be combined with previous ones in BLC, in 3TX of a 802.16e MIMO-HARQ scheme.

In order to verify the superiority of the proposed scheme, the simulation based on a low correlation MIMO model [WiMAX ,2007] with *Mt r M 2* is undertaken here. In particular, we show two examples of comparison in this paper: *VSTBC in SLC* v.s. *SFBC in BLC* and *SFBC in BLC* v.s. *SM in BLC*. The delay profile of each path is evaluated under ITU-R [ITU-R ,2000] Pedestrian Type-B 3km/hr (PB3) or Vehicular Type-A 60km/hr (VA60). Furthermore, PUSC with 10Mhz bandwidth is assumed and the coding scheme is based on 802.16e CTC. In addition, we postulate that the receiver has perfect CSI. The HARQ round trip interval is 10ms, and the subpacket will be shifted by 3 subchannel length to gain higher diversity in frequency domain. The subpacket size for each coding rate is summarized in Table 3, where *NEP* is the number of information bits before feeding into FEC encoder. Note that we concentrate on CC mode and the comparison with IR mode is beyond our scope due

modulation coding scheme *NEP*

Table 3. *NEP* size of different code rate with CC

**Example 1: VSTBC in SLC v.s. SFBC in BLC.** 

of a 802.16e MIMO-HARQ scheme.

QPSK 1*/*2 480 QPSK 3*/*4 432 16QAM 1*/*2 480 16QAM 3*/*4 432 64QAM 1*/*2 432 64QAM 2*/*3 384 64QAM 3*/*4 432 64QAM 5*/*6 480

We focus on simulation results of PB3 and VA60. In 1TX, the receiver is a soft MMSE detector for both cases, and in 2TX the VSTBC-MRC or SFBC-MRC is used. Simulation results of 2TX packet error rate (PER) in Figure 9 and 10 show that our proposed scheme has poorer error performance in low mobility. However, the interference terms in equation 9 will impact the performance of VSTBC in SLC scheme as the mobility increases. Hence, the proposed scheme becomes superior in moderate and higher speed scenarios. Nevertheless, a Doppler estimator is generally not available at the receiver; hence conventional 802.16e cannot be guaranteed to be superior at all mobility levels. The PER in 3TX is not evaluated, because the retransmitted subpacket cannot be combined with previous ones in BLC, in 3TX

**6. Numerical results** 

to page restrictions.

Fig. 9. VSTBC SLC v.s. SFBC BLC in 2TX PB3 CC

Fig. 10. VSTBC SLC v.s. SFBC BLC in 2TX VA60 CC

Hybrid ARQ Utilizing Lower Rate Retransmission over MIMO Wireless Systems 145

throughput comparison in 13 and 14. Thus, when the channel model is PB3 with low spatial correlation, it is shown that the proposed scheme always outperforms the ones in 802.16e, especially in higher coding rate scenarios. In terms of throughput, the proposed scheme can achieve higher throughput in low signal to noise ratio (SNR) region with the same coding rate. Nonetheless, the throughput curves are similar in high SNR region. The results also

show that the proposed scheme is less sensitive to improper link adaptation.

Fig. 13. Soft MMSE BLC v.s. SFBC MRC BLC in PB3 CC

Fig. 14. LSD BLC v.s. SFBC MRC BLC in PB3 CC

#### **Example 2: SFBC in BLC v.s. SM in BLC.**

1TX, the SM detection is the same for soft MMSE/LSD for 802.16e and the proposed scheme. For later retransmissions than 1TX, the detection method is soft MMSE/LSD in 802.16e and SFBC-MRC in the proposed scheme. The maximum number of retransmission assumed in our simulation is 3. We first show that the PER results in 11 and 12, and it then follows by

Fig. 11. Soft MMSE BLC v.s. SFBC MRC BLC in 3TX PB3 CC

Fig. 12. LSD BLC v.s. SFBC MRC BLC in 3TX PB3 CC

1TX, the SM detection is the same for soft MMSE/LSD for 802.16e and the proposed scheme. For later retransmissions than 1TX, the detection method is soft MMSE/LSD in 802.16e and SFBC-MRC in the proposed scheme. The maximum number of retransmission assumed in our simulation is 3. We first show that the PER results in 11 and 12, and it then follows by

**Example 2: SFBC in BLC v.s. SM in BLC.** 

Fig. 11. Soft MMSE BLC v.s. SFBC MRC BLC in 3TX PB3 CC

Fig. 12. LSD BLC v.s. SFBC MRC BLC in 3TX PB3 CC

throughput comparison in 13 and 14. Thus, when the channel model is PB3 with low spatial correlation, it is shown that the proposed scheme always outperforms the ones in 802.16e, especially in higher coding rate scenarios. In terms of throughput, the proposed scheme can achieve higher throughput in low signal to noise ratio (SNR) region with the same coding rate. Nonetheless, the throughput curves are similar in high SNR region. The results also show that the proposed scheme is less sensitive to improper link adaptation.

Fig. 13. Soft MMSE BLC v.s. SFBC MRC BLC in PB3 CC

Fig. 14. LSD BLC v.s. SFBC MRC BLC in PB3 CC

**8** 

*Czech Republic* 

**On Efficiency of ARQ and HARQ Entities** 

*Czech Technical University in Prague, Faculty of Electrical Engineering,* 

During data exchange in wireless networks, an error in transmission can occur. Corrupted data cannot be further processed without a correction. A technique based on either Automatic Repeat reQuest (ARQ) or Forward Error Correction (FEC) is conventionally used to repair erroneous data in wireless networks. The ARQ is backward mechanism that uses a feedback channel for the confirmation of error-free data delivery or to request a retransmission of corrupted data. This method can increase a network throughput if radio channel conditions are getting worse (Sambale et al., 2008). On the other hand, the ARQ method increases the delay of packets due to the retransmission of former unsuccessfully received packets. The FEC can increase user's data throughput over the channel with poor quality despite the fact that additional redundant bits are coded together with users' data at the transmitter side. The method combining both above mentioned methods is called Hybrid ARQ (HARQ). All three error correction mechanisms are implemented on physical

The performance of ARQ defined in standard IEEE 802.16e (IEEE802.16e, 2006) depends on the setting of several parameters such as size of user data carried in a frame, size of ARQ block, size of PDU (Protocol Data Unit), limit of retransmission timeout timer, or type of packet acknowledgement (Lee & Choi, 2008). Evaluation of the type of packet acknowledgment for different channel condition is presented in (Kang & Jang, 2008). In other paper, the authors evaluate the ARQ performance for different ARQ parameters (Tykhomyrov et al., 2007). This work is later on enhanced by analysis of the impact of PDU size on IEEE 802.16e networks performance while ARQ mechanism is used (Martikainen et al., 2008). Further, a comparison of ARQ and HARQ performance in IEEE 802.16 networks is presented in (Sayenko et al., 2008). This paper also compares the amount of overhead generated by ARQ and HARQ. The optimal PDU size and MAC overhead due to the packets retransmission is analyzed in (Hoymann, 2005). The authors in (Sengupta et al., 2005) propose to adjust the MAC PDU size depending on the channel state to achieve the best ARQ performance. The paper is extended for analysis of a combination of error correction techniques such as ARQ, FEC or MAC PDU aggregation on the VoIP speech quality (Sengupta et al., 2008). Authors proof the improvement of VoIP speech quality by using these techniques. In (Chen & De Marca, 2008), the authors investigate an optimization

**1. Introduction** 

and/or Medium Access Control (MAC) layer.

of ARQ parameter setting from the link throughput point of view.

**Interaction in WiMAX Networks** 

Zdenek Becvar and Pavel Mach

#### **7. Conclusions**

In this chapter, we proposed a novel scheme to gain better performance, which reduces the receiver complexity (by 50% ∼ 70% in 2 × 2 MIMO scenario) as well as the buffer requirement. From the simulation, it is observable that the proposed scheme can achieve better performance than conventional 802.16e schemes in scenarios with moderate or high mobility. In another set of simulations, we have compared the proposed scheme with the BLC in 802.16e architecture. The results have verified that our scheme is less sensitive to inappropriate link adaptation.

#### **8. References**


### **On Efficiency of ARQ and HARQ Entities Interaction in WiMAX Networks**

Zdenek Becvar and Pavel Mach

*Czech Technical University in Prague, Faculty of Electrical Engineering, Czech Republic* 

#### **1. Introduction**

146 Advanced Transmission Techniques in WiMAX

In this chapter, we proposed a novel scheme to gain better performance, which reduces the

×

requirement. From the simulation, it is observable that the proposed scheme can achieve better performance than conventional 802.16e schemes in scenarios with moderate or high mobility. In another set of simulations, we have compared the proposed scheme with the BLC in 802.16e architecture. The results have verified that our scheme is less sensitive to

Alamouti, S. M. (1998). A simple transmit diversity technique for wireless communications,

Chase, D. (1985). Code combining- a maximum-likelihood decoding approach for

Chen, C.M.; Hsu, J.Y.; Kuo, P.H.& Ting, P.A. (2009). MIMO Hybrid-ARQ Utilizing Lower

Damen, M. O.; Gamal, H. El & Caire G. (2003). On maximum-likelihood detection and the

Foschini, G. (1996). Layered space-time architecture for wireless communication in a fading

Gao, Y.; Li, G.; Chen, K. & Wu, X. (2007) A Novel HARQ Scheme Utilizing the Iterative Soft-

IEEE (2007). IEEE 802.16e-2005, IEEE Standard for Local and Metropolitan Area Networks, Part 16: Air interface for Fixed Broadband Wireless Access Systems, Oct. 2007. ITU-R (2000). Recommendation ITU-R M.1225 Guidelines for Evaluation of Radio

Jia, M. L.; Kuang J. M. & He, Z. W. (2006). Enhanced HARQ Employing LDPC Coded

Kaiser, S. (2003). Space frequency block codes and code division multiplexing in OFDM

Lee, I. & Sundberg, C.-E.W. (2007). Reduced-Complexity Receiver Structures for Space–Time

Lin, S.; Costello, Jr., D. J. & Miller, M. J. (1984). Automatic-Repeat Request Error-Control

WiMAX (2007). WiMAX Forum Mobile Radio Conformance Tests (MRCT) Release 1.0

Schemes, *IEEE Communication Magazine*, vol. 22, pp5-17, Dec. 1984.

*IEEE Journal on Selected Areas in Communications*, vol. 16, no. 8, pp. 1451– 1458, Oct.

combining an arbitrary number of noisy packets, *IEEE Transaction Communications*,

Rate Retransmission over Mobile WiMAX System, IEEE Mobile WiMAX, pp. 129 -

search for the closest lattice point, *IEEE Transactions on Information Theory,* vol. 49,

environment when using multielement antennas, *Bell Labs Technical Journal,* vol. 1,

information Feedback in MIMO System *Vehicular Technology Conference*, pp. 423-

Constellation Rearrangement with 64QAM. *International Conference on* 

systems, IEEE *Global Telecommunications Conference,* vol. 4, 1-5, pp. 2360-2364, Dec.

Bit-Interleaved Coded Modulation Systems, *IEEE Transactions on Communications,*

2 MIMO scenario) as well as the buffer

∼

70% in 2

**7. Conclusions** 

**8. References** 

1998.

receiver complexity (by 50%

inappropriate link adaptation.

vol. 33, pp. 385-393, 1985.

Issue 10, pp. 2389 - 2402, Oct. 2003.

Transmission Technologies for IMT-2000

vol. 55, Issue 1, pp. 142-150, Jan. 2007.

*Communication Technology*, pp. 1-4, Nov. 2006.

Approved Specification, Revision 2.0.0, Dec. 2007.

134 , July 2009.

no. 2, pp. 41–59, 1996.

424, Apr. 2007.

2003.

During data exchange in wireless networks, an error in transmission can occur. Corrupted data cannot be further processed without a correction. A technique based on either Automatic Repeat reQuest (ARQ) or Forward Error Correction (FEC) is conventionally used to repair erroneous data in wireless networks. The ARQ is backward mechanism that uses a feedback channel for the confirmation of error-free data delivery or to request a retransmission of corrupted data. This method can increase a network throughput if radio channel conditions are getting worse (Sambale et al., 2008). On the other hand, the ARQ method increases the delay of packets due to the retransmission of former unsuccessfully received packets. The FEC can increase user's data throughput over the channel with poor quality despite the fact that additional redundant bits are coded together with users' data at the transmitter side. The method combining both above mentioned methods is called Hybrid ARQ (HARQ). All three error correction mechanisms are implemented on physical and/or Medium Access Control (MAC) layer.

The performance of ARQ defined in standard IEEE 802.16e (IEEE802.16e, 2006) depends on the setting of several parameters such as size of user data carried in a frame, size of ARQ block, size of PDU (Protocol Data Unit), limit of retransmission timeout timer, or type of packet acknowledgement (Lee & Choi, 2008). Evaluation of the type of packet acknowledgment for different channel condition is presented in (Kang & Jang, 2008). In other paper, the authors evaluate the ARQ performance for different ARQ parameters (Tykhomyrov et al., 2007). This work is later on enhanced by analysis of the impact of PDU size on IEEE 802.16e networks performance while ARQ mechanism is used (Martikainen et al., 2008). Further, a comparison of ARQ and HARQ performance in IEEE 802.16 networks is presented in (Sayenko et al., 2008). This paper also compares the amount of overhead generated by ARQ and HARQ. The optimal PDU size and MAC overhead due to the packets retransmission is analyzed in (Hoymann, 2005). The authors in (Sengupta et al., 2005) propose to adjust the MAC PDU size depending on the channel state to achieve the best ARQ performance. The paper is extended for analysis of a combination of error correction techniques such as ARQ, FEC or MAC PDU aggregation on the VoIP speech quality (Sengupta et al., 2008). Authors proof the improvement of VoIP speech quality by using these techniques. In (Chen & De Marca, 2008), the authors investigate an optimization of ARQ parameter setting from the link throughput point of view.

On Efficiency of ARQ and HARQ Entities Interaction in WiMAX Networks 149

In WiMAX, each data burst generated either by the MS or the BS is segmented into PDUs. These PDUs are further mapped into MAC frame. A PDU usually consists several blocks

> *i,k data i,k ARQ blocks*

where *i,k Sdata* is a total size of data of *i-th* user in *k-th* frame, and similarly *i,k SARQ blocks* represents a block size defined by parameter denoted in the standard as ARQ\_Block\_Size (IEEE802.16e, 2006). This parameter is carried in TLV (Type/Length/Value) section of registration messages (REG-REQ/RSP) exchanged between the BS and MS (see (IEEE802.16e, 2006)). The parameter ARQ\_Block\_Size can take values from the following range: 16, 32, 64, 128, 256, 512 and 1024 bytes. During a transmission, a sequence of consecutive blocks is sent in the PDU. After that the receiver evaluates whether the data are received correctly or not and sends an appropriate feedback message to the transmitter. Note that all transmitted blocks (*Nblock*) have to be confirmed by ACK or NACK even if all blocks are received without errors. The IEEE 802.16e standard defines four types of acknowledgments: Selective ACK entry, Cumulative ACK entry, Cumulative with Selective

The first type of acknowledgment uses selective maps to provide feedback to the transmitter. In the selective map, each bit corresponds to one ARQ block. A bit set to "1" indicates error-free reception of the corresponding ARQ block. The second type, Cumulative ACK entry, is based on the utilization of sequence maps. A sequence map defines a group of consecutive blocks where each group includes a sequence of only erroneous blocks or sequence of only error free blocks. The sequence maps can contain two or three sequences with a length of 64 or 16 blocks respectively. The third type of ACK combines the previous two types. Finally, the last type combines the second type with ability to acknowledge ARQ

The ACK or NACK is sent through above mentioned feedback message. The feedback is transmitted in the next frame after the data transmission. The feedback message contains 8 bit field indicating Message ID and the rest of the message is dedicated to field consisting ARQ\_Feedback\_Payload. The ARQ payload can be carried either via standalone ARQ feedback message or by piggybacking the ARQ payload to the user's data block. The payload is always carried in a single PDU. The ARQ\_Feedback\_Payload includes one or

more ARQ\_Feedback\_IE (see Table 1) where IE stands for an Information Element.

*S*

*<sup>S</sup> <sup>N</sup>*

*i,k*

(1)

*blocks*

Fig. 1. Principle of conventional ARQ

*Nblock*, which number is given by following equation:

ACK entry and Cumulative with Block Sequence ACK entry.

blocks in the form of block sequences.

In conventional WiMAX network, the ARQ and HARQ work independently on each other (IEEE802.16e, 2006). In standalone ARQ process, a number of blocks received with errors increases as the link quality between the transmitter and the receiver decreases. Thus, if the Block Error Rate (BLER) is more significant, the amount of retransmitted blocks is higher as well. It can be assumed that if the channel quality is high, most of the blocks are transferred without errors and the number of unsuccessfully received blocks is kept to minimum. In such case, the transmission of positive acknowledgement (ACK) of correctly delivered blocks appears more often than negative acknowledgement (NACK) of corrupted blocks. This assumption is considered in (Becvar & Bestak, 2011), where authors propose to send only NACKs to significantly reduce signaling overhead introduced by ARQ mechanism.

On the other hand, the HARQ is able to detect and correct the most of the radio channel errors. However, due to the limitation of a number of retransmissions, some data may not be delivered without errors if only HARQ is utilized. Consequently, these data have to be retransmitted by ARQ process. The conventional ARQ has to acknowledge all data independently on the result of HARQ procedure. In order to significantly reduce signaling overhead, an interaction of both ARQ and HARQ methods should be utilized, see, e.g., (Maheshwari et al., 2008)).

The contributions of this chapter are as follows. Firstly, the results of improved ARQ scheme according to (Becvar & Bestak, 2011) cooperating with HARQ is compared to the results achieved by the conventional ARQ scheme with enabled and disabled cooperation between both entities. Secondly, while only one hop communication is assumed when data are sent only between a mobile station (MS) and a base station (BS) in (Becvar & Bestak, 2011), this chapter analyzes the impact of relay stations (RS), defined in IEEE 802.16j (IEEE802.16j, 2009), on the performance of individual methods. The extended simulations are performed considering various setting of parameters. The amount of generated overhead is the metric for the performance assessment.

The rest of this chapter is organized as follows. In the next section, the principle of ARQ and HARQ used in WiMAX are described. In addition, the optimization of conventional ARQ scheme according to (Becvar & Bestak, 2011) is presented in this section. In the section 3, the overhead of HARQ algorithm in WiMAX networks is evaluated. The section 4 provides an overview on simulation model and contemplates the parameters applied in simulator. The section 5 presents the simulation results. Last section gives our conclusions.

#### **2. ARQ and HAQR in WiMAX**

This section provides overview on conventional ARQ and HARQ used in WiMAX networks. Further, the innovative ARQ proposed in (Becvar & Bestak, 2011) is also described to enable easy understanding of results presented in next sections.

#### **2.1 Conventional ARQ**

The principle of conventional ARQ method according to the IEEE 802.16e standard and the structure of user's information carried in the frame are depicted in Fig. 1.

Fig. 1. Principle of conventional ARQ

In conventional WiMAX network, the ARQ and HARQ work independently on each other (IEEE802.16e, 2006). In standalone ARQ process, a number of blocks received with errors increases as the link quality between the transmitter and the receiver decreases. Thus, if the Block Error Rate (BLER) is more significant, the amount of retransmitted blocks is higher as well. It can be assumed that if the channel quality is high, most of the blocks are transferred without errors and the number of unsuccessfully received blocks is kept to minimum. In such case, the transmission of positive acknowledgement (ACK) of correctly delivered blocks appears more often than negative acknowledgement (NACK) of corrupted blocks. This assumption is considered in (Becvar & Bestak, 2011), where authors propose to send only NACKs to significantly reduce signaling overhead

On the other hand, the HARQ is able to detect and correct the most of the radio channel errors. However, due to the limitation of a number of retransmissions, some data may not be delivered without errors if only HARQ is utilized. Consequently, these data have to be retransmitted by ARQ process. The conventional ARQ has to acknowledge all data independently on the result of HARQ procedure. In order to significantly reduce signaling overhead, an interaction of both ARQ and HARQ methods should be utilized, see, e.g.,

The contributions of this chapter are as follows. Firstly, the results of improved ARQ scheme according to (Becvar & Bestak, 2011) cooperating with HARQ is compared to the results achieved by the conventional ARQ scheme with enabled and disabled cooperation between both entities. Secondly, while only one hop communication is assumed when data are sent only between a mobile station (MS) and a base station (BS) in (Becvar & Bestak, 2011), this chapter analyzes the impact of relay stations (RS), defined in IEEE 802.16j (IEEE802.16j, 2009), on the performance of individual methods. The extended simulations are performed considering various setting of parameters. The amount of generated overhead is the metric

The rest of this chapter is organized as follows. In the next section, the principle of ARQ and HARQ used in WiMAX are described. In addition, the optimization of conventional ARQ scheme according to (Becvar & Bestak, 2011) is presented in this section. In the section 3, the overhead of HARQ algorithm in WiMAX networks is evaluated. The section 4 provides an overview on simulation model and contemplates the parameters applied in simulator. The

This section provides overview on conventional ARQ and HARQ used in WiMAX networks. Further, the innovative ARQ proposed in (Becvar & Bestak, 2011) is also

The principle of conventional ARQ method according to the IEEE 802.16e standard and the

section 5 presents the simulation results. Last section gives our conclusions.

described to enable easy understanding of results presented in next sections.

structure of user's information carried in the frame are depicted in Fig. 1.

introduced by ARQ mechanism.

(Maheshwari et al., 2008)).

for the performance assessment.

**2. ARQ and HAQR in WiMAX** 

**2.1 Conventional ARQ** 

In WiMAX, each data burst generated either by the MS or the BS is segmented into PDUs. These PDUs are further mapped into MAC frame. A PDU usually consists several blocks *Nblock*, which number is given by following equation:

$$\mathbf{N}\_{\text{blocks}}^{i,k} = \frac{S\_{\text{data}}^{i,k}}{S\_{\text{ARQ}-\text{blocks}}^{i,k}} \tag{1}$$

where *i,k Sdata* is a total size of data of *i-th* user in *k-th* frame, and similarly *i,k SARQ blocks* represents a block size defined by parameter denoted in the standard as ARQ\_Block\_Size (IEEE802.16e, 2006). This parameter is carried in TLV (Type/Length/Value) section of registration messages (REG-REQ/RSP) exchanged between the BS and MS (see (IEEE802.16e, 2006)). The parameter ARQ\_Block\_Size can take values from the following range: 16, 32, 64, 128, 256, 512 and 1024 bytes. During a transmission, a sequence of consecutive blocks is sent in the PDU. After that the receiver evaluates whether the data are received correctly or not and sends an appropriate feedback message to the transmitter. Note that all transmitted blocks (*Nblock*) have to be confirmed by ACK or NACK even if all blocks are received without errors. The IEEE 802.16e standard defines four types of acknowledgments: Selective ACK entry, Cumulative ACK entry, Cumulative with Selective ACK entry and Cumulative with Block Sequence ACK entry.

The first type of acknowledgment uses selective maps to provide feedback to the transmitter. In the selective map, each bit corresponds to one ARQ block. A bit set to "1" indicates error-free reception of the corresponding ARQ block. The second type, Cumulative ACK entry, is based on the utilization of sequence maps. A sequence map defines a group of consecutive blocks where each group includes a sequence of only erroneous blocks or sequence of only error free blocks. The sequence maps can contain two or three sequences with a length of 64 or 16 blocks respectively. The third type of ACK combines the previous two types. Finally, the last type combines the second type with ability to acknowledge ARQ blocks in the form of block sequences.

The ACK or NACK is sent through above mentioned feedback message. The feedback is transmitted in the next frame after the data transmission. The feedback message contains 8 bit field indicating Message ID and the rest of the message is dedicated to field consisting ARQ\_Feedback\_Payload. The ARQ payload can be carried either via standalone ARQ feedback message or by piggybacking the ARQ payload to the user's data block. The payload is always carried in a single PDU. The ARQ\_Feedback\_Payload includes one or more ARQ\_Feedback\_IE (see Table 1) where IE stands for an Information Element.

On Efficiency of ARQ and HARQ Entities Interaction in WiMAX Networks 151

(burst #5 in Fig. 1) can be delayed by one frame. It causes a delay of retransmitted packets with duration that corresponds to at least 3 times of frame duration (e.g., if the frame

The new and retransmitted data are sent within the same frame only if the requested capacity (new data plus retransmitted data) is available. The WiMAX technology implements Stop-and-Wait mechanism that requests a confirmation of the previous block before transmitting subsequent blocks. The number of blocks that can be unconfirmed before a transmission of the consequent blocks is defined in the standard by the parameter

The innovative ARQ takes into consideration that the number of blocks received with errors increases as the link quality between transmitter and receiver decreases. This scheme adaptively selects one of three different ways of data delivery confirmation: i) conventional ARQ, ii) transmission of only NACK (ARQ Scheme I), iii) retransmission of only corrupted

The first type of data acknowledgement, the conventional ARQ, was already explained

The second type (ARQ Scheme I) assumes ARQ feedback message and ARQ\_Feedback\_IEs with the same structure as the conventional IEEE 802.16 ARQ feedback message. However in this proposal, the ARQ feedback is sent only if a received PDU contains at least one erroneous block. If all blocks in the PDU are error free, no feedback is sent. The PDU is assumed to be correctly transferred if the transmitter receives no feedback in the following W frames after the transmission. If the feedback with NACK is not delivered, the data conveying the delay sensitive services (e.g., VoIP) are assumed to be lost since the delay caused by repeated ARQ retransmission is significant. In case of services not sensitive to delay, data belonging to lost NACK can be retransmitted using upper layer protocols, e.g., TCP (Transmission Control Protocol). As the probability of lost packet or packet with errors together with the NACK feedback is very low, the increase of overhead due to upper layer

The third way of data acknowledgement (ARQ Scheme II) is based on the same assumptions as the previous one. The ACK feedback is likewise transmitted only if there is at least one block with errors. A block is assumed to be error-free if no feedback is received in one of the

The overhead generated by the innovative ARQ (denoted as ARQ PIII) by a user in one

*IE*

where *NIE* is the number of IEs carried in one ARQ feedback message, *MNIE* corresponds to the number of ACK maps in ARQ\_Feedback\_IE, *B* stands for the number of BSNs included in one message and *res* is the number of bits used for an alignment of the feedback message

*N SizeARQ \_ FB\_ III 8 18 min 16 16 M ,10 B 11 res <sup>N</sup>*

*IE*

(5)

following W frames after the transmission of appropriate data frame.

frame can be calculated according to the following equation:

duration is 10 ms, the packet delay is at least 30 ms).

ARQ\_Window\_Size.

**2.2 Innovative ARQ** 

blocks (ARQ Scheme II).

protocols is negligible.

before.


Table 1. Structure of ARQ\_Feedback\_IE (IEEE802.16e, 2006)

The size of an IE of each ARQ feedback message can be calculated according to equation:

$$\text{Size}\_{\text{ARQ\\_FB\\_IE}}\text{ [bits]} = \text{32} + \text{(M} \times \text{16)} \tag{2}$$

where *M* represents the number of maps carried in one ARQ\_Feedback\_IE (see Table 1). Consequently, the overall size of whole feedback message is given by following formula:

$$\text{Size}\_{ARQ\\_FB} \text{ [bits]} = \\$ + \sum\_{i=1}^{N\_{IE}} \text{Size}\_{ARQ\\_FB\\_IE\_i} \tag{3}$$

where *NIE* corresponds to the amount of information elements carried in one ARQ Feedback message and the first eight bits represents the ARQ feedback message overhead (i.e., Message ID field). The overhead transmitted in all considered frames (*Nframe*) is equal to the sum of partial overheads over the *Nframe*:

$$\{OH\_{ConvARQ} \mid \text{bits} \} = \sum\_{i=1}^{N\_{form}} \text{Size}\_{ARQ\\_FB\_i} \tag{4}$$

As indicated in Fig. 1, the retransmission of erroneous blocks cannot be accomplished before the third frame after the original transmission since the transmitter receives NACK in the next frame after transmission (2nd frame). Hence a request for additional resources can be created earliest at the upcomming frame (3rd frame). Therefore, the dedicated resources are not available before the 4th frame. The retransmission of data (burst #2 in Fig. 1) can be scheduled either together with normally ordered data (burst #5 in Fig. 1) or the new data (burst #5 in Fig. 1) can be delayed by one frame. It causes a delay of retransmitted packets with duration that corresponds to at least 3 times of frame duration (e.g., if the frame duration is 10 ms, the packet delay is at least 30 ms).

The new and retransmitted data are sent within the same frame only if the requested capacity (new data plus retransmitted data) is available. The WiMAX technology implements Stop-and-Wait mechanism that requests a confirmation of the previous block before transmitting subsequent blocks. The number of blocks that can be unconfirmed before a transmission of the consequent blocks is defined in the standard by the parameter ARQ\_Window\_Size.

#### **2.2 Innovative ARQ**

150 Advanced Transmission Techniques in WiMAX

0x0...Selective ACK 0x1...Cumulative ACK

Selective (16 blocks) or

The size of an IE of each ARQ feedback message can be calculated according to equation:

where *M* represents the number of maps carried in one ARQ\_Feedback\_IE (see Table 1). Consequently, the overall size of whole feedback message is given by following formula:

where *NIE* corresponds to the amount of information elements carried in one ARQ Feedback message and the first eight bits represents the ARQ feedback message overhead (i.e., Message ID field). The overhead transmitted in all considered frames (*Nframe*) is equal to the

*Size [bits] 8 Size*

*OH [bits] Size*

*IE*

*frame*

*N ConvARQ ARQ \_ FB i 1*

As indicated in Fig. 1, the retransmission of erroneous blocks cannot be accomplished before the third frame after the original transmission since the transmitter receives NACK in the next frame after transmission (2nd frame). Hence a request for additional resources can be created earliest at the upcomming frame (3rd frame). Therefore, the dedicated resources are not available before the 4th frame. The retransmission of data (burst #2 in Fig. 1) can be scheduled either together with normally ordered data (burst #5 in Fig. 1) or the new data

*N ARQ \_ FB ARQ \_ FB\_ IE i 1*

0x2...Cumulative with Selective

0x3...Cumulative with Block Sequence

Cumulative maps (2 x 64 blocks / 3 x 16 blocks) Cumulative maps: 1 bit sequence format (2 or 3 blocks), 2/3bits Sequence ACK (ACK/NACK of sequence), (2x6) / (3x4) bits Sequence length

*Size [bits] 32 M 16 ARQ \_ FB\_ IE* (2)

*i*

*i*

(3)

(4)

Last 1 bit Identify the last IE in ARQ\_Feedback

BSN 11 bits Block Sequence Number (0...2047) Number of ACK Map 2 bits Number of Maps (M) = 1,2,3 or 4

**Syntax Size Notes** 

ACK Type 2 bits

Maps M x 16

sum of partial overheads over the *Nframe*:

CID 16 bits Connection ID

bits

Table 1. Structure of ARQ\_Feedback\_IE (IEEE802.16e, 2006)

The innovative ARQ takes into consideration that the number of blocks received with errors increases as the link quality between transmitter and receiver decreases. This scheme adaptively selects one of three different ways of data delivery confirmation: i) conventional ARQ, ii) transmission of only NACK (ARQ Scheme I), iii) retransmission of only corrupted blocks (ARQ Scheme II).

The first type of data acknowledgement, the conventional ARQ, was already explained before.

The second type (ARQ Scheme I) assumes ARQ feedback message and ARQ\_Feedback\_IEs with the same structure as the conventional IEEE 802.16 ARQ feedback message. However in this proposal, the ARQ feedback is sent only if a received PDU contains at least one erroneous block. If all blocks in the PDU are error free, no feedback is sent. The PDU is assumed to be correctly transferred if the transmitter receives no feedback in the following W frames after the transmission. If the feedback with NACK is not delivered, the data conveying the delay sensitive services (e.g., VoIP) are assumed to be lost since the delay caused by repeated ARQ retransmission is significant. In case of services not sensitive to delay, data belonging to lost NACK can be retransmitted using upper layer protocols, e.g., TCP (Transmission Control Protocol). As the probability of lost packet or packet with errors together with the NACK feedback is very low, the increase of overhead due to upper layer protocols is negligible.

The third way of data acknowledgement (ARQ Scheme II) is based on the same assumptions as the previous one. The ACK feedback is likewise transmitted only if there is at least one block with errors. A block is assumed to be error-free if no feedback is received in one of the following W frames after the transmission of appropriate data frame.

The overhead generated by the innovative ARQ (denoted as ARQ PIII) by a user in one frame can be calculated according to the following equation:

$$Size\_{ARQ\\_FB\\_III} = 8 + 18 + \min\left\{\sum^{N\_E} 16 + 16 \times M\_{N\_E}, 10 + B \times 11\right\} + \text{res} \tag{5}$$

where *NIE* is the number of IEs carried in one ARQ feedback message, *MNIE* corresponds to the number of ACK maps in ARQ\_Feedback\_IE, *B* stands for the number of BSNs included in one message and *res* is the number of bits used for an alignment of the feedback message

On Efficiency of ARQ and HARQ Entities Interaction in WiMAX Networks 153

2006)). In the simulations performed in this chapter, the *Region ID* is not considered. The actual amount of bits of *SubB* depends on the HARQ Type. Based on the (IEEE802.16e,

> *SubB 8 N (RCID 20 DIUC ) SubB 8 N (RCID 20) SubB 8 N (R 22 DIUC )*

where *Nsub* is a number of sub-bursts; *RCID* represents a size of Reduced CID; and *DIUC*

For the case when a low number of bursts are transmitted within a frame, the utilization of so called Compact HARQ DL/UL maps enables to reduce an overhead (see (IEEE802.16e, 2006)). The overhead generated by compact version of maps is not dependent on the HARQ

 

(8)

(9)

2006), the size of message according to the sub-bursts is following:

*CC sub*

 

type. The amount of overhead can be expressed by the next equations:

included and 4 bits if the information are not included).

Reduced CID.

and UL *(b)*

*DL*

*UL HARQOHcomp 12 RCID HCI CCI HARQOHcomp 12 RCID HCI*

where *HCI* is a size of HARQ control IE (8 bits if HARQ is enabled and 4 bits if HARQ is temporary disabled); CCI is a size of CQICH control IE (16 bits if CQICH information are

The simple evaluation of equations for full and compact HARQ maps enables to determine which kind of maps generates minimum management overhead over the number of HARQ sub-bursts (see Fig. 2). As the results show, the compact version of maps is profitable for all numbers of sub-bursts in UL as well as for up to 12 sub-bursts in DL over all length of

 (a) (b) Fig. 2. Comparison of the overhead generated by compact and full HARQ maps for DL *(a)* 

*IR CTC sub IR CC sub*

represents the size of optional field, denoted as DIUC, containing 8 bits if included.

length to integer number of bytes. The overhead generated by new ARQ scheme is given by the following equitation:

$$OH\_{Schmelill} = \sum^{N\_{fmm}} Size\_{ARQ\\_FB\\_III\_{N\_{fmm}}} \tag{6}$$

#### **2.3 HARQ**

The utilization of ARQ with support of FEC is known as HARQ. The HARQ method uses not only retransmitted packets to reconstruct the original error free packets, but it also utilizes the packets received with errors. The original packet can be reconstructed by a combination of several versions of packet with errors. The HARQ described in (IEEE802.16e, 2006) uses two different types of reconstruction: Chase Combining (CC) and Incremental Redundancy (IR).

The first version of HARQ is denoted as Type I HARQ Chase Combining. In this case, blocks of data together with a CRC code are encoded using a FEC coder before transmission. If the channel quality is low and errors of data are identified, the data block is not discarded however it is kept in the memory. In the next phase, the receiver requests for retransmission of this data block. The retransmitted block of data is then combined with the previous blocks received with errors. Combining more versions of the data blocks improves the probability of correct decoding even if all of them are received with errors.

Optionally, the IEEE802.16 standard also supports type II HARQ, which is known as Incremental Redundancy. In case of IR HARQ, the FEC coder codes one packet into several subpackets. Each of subpacket is coded with different code ratio. The subpackets are distinguished by 2-bits SubPacket IDentifier (SPID). If the packet is transmitted for the first time, the subpacket with SPID=00 is sent. The successful receive of the packet at the destination station is indicated by ACK. Otherwise, the transmitter sends a NACK and the transmitter has to send another packet carrying one of four subpackets. Both received packet (the first transmission and retransmissions) are again combined by receiver to increase the probability of correct decoding.

The overhead introduced by HARQ in WiMAX depends on the HARQ Type as follows. Firstly, the acknowledgment of HARQ bursts by modification of AI\_SN (HARQ Identifier Sequence Number) of appropriate ACID (HARQ Channel ID) is assumed (for more information, see (IEEE802.16e, 2006)). The AI\_SN is included in HARQ DL or UL. The size of HARQ map can be described by the subsequent formula:

$$\begin{aligned} \text{HARQOH}\_{DL} &= \begin{cases} 64 + SubB + res \dots \text{Re}\,\text{ign}\,\text{ID\\_ON} \\ 40 + SubB + res \dots \text{Re}\,\text{ign}\,\text{ID\\_OFF} \end{cases} \\\\ \text{HARQOH}\_{UL} &= \begin{cases} 48 + SubB + res \dots \text{Re}\,\text{ign}\,\text{ID\\_ON} \\ 24 + SubB + res \dots \text{Re}\,\text{ign}\,\text{ID\\_OFF} \end{cases} \end{aligned} \tag{7}$$

where *SubB* is a size of management overhead according to a sub-burst. The amount of management overhead also depends on the utilization of *Region ID* (see (IEEE802.16e,

length to integer number of bytes. The overhead generated by new ARQ scheme is given by

The utilization of ARQ with support of FEC is known as HARQ. The HARQ method uses not only retransmitted packets to reconstruct the original error free packets, but it also utilizes the packets received with errors. The original packet can be reconstructed by a combination of several versions of packet with errors. The HARQ described in (IEEE802.16e, 2006) uses two different types of reconstruction: Chase Combining (CC) and Incremental

The first version of HARQ is denoted as Type I HARQ Chase Combining. In this case, blocks of data together with a CRC code are encoded using a FEC coder before transmission. If the channel quality is low and errors of data are identified, the data block is not discarded however it is kept in the memory. In the next phase, the receiver requests for retransmission of this data block. The retransmitted block of data is then combined with the previous blocks received with errors. Combining more versions of the data blocks improves the probability

Optionally, the IEEE802.16 standard also supports type II HARQ, which is known as Incremental Redundancy. In case of IR HARQ, the FEC coder codes one packet into several subpackets. Each of subpacket is coded with different code ratio. The subpackets are distinguished by 2-bits SubPacket IDentifier (SPID). If the packet is transmitted for the first time, the subpacket with SPID=00 is sent. The successful receive of the packet at the destination station is indicated by ACK. Otherwise, the transmitter sends a NACK and the transmitter has to send another packet carrying one of four subpackets. Both received packet (the first transmission and retransmissions) are again combined by receiver to increase the

The overhead introduced by HARQ in WiMAX depends on the HARQ Type as follows. Firstly, the acknowledgment of HARQ bursts by modification of AI\_SN (HARQ Identifier Sequence Number) of appropriate ACID (HARQ Channel ID) is assumed (for more information, see (IEEE802.16e, 2006)). The AI\_SN is included in HARQ DL or UL. The size

*64 SubB res ...Re gionID\_ON HARQOH 40 SubB res...Re gionID\_OFF*

*48 SubB res ...Re gionID\_ON HARQOH 24 SubB res...Re gionID\_OFF*

where *SubB* is a size of management overhead according to a sub-burst. The amount of management overhead also depends on the utilization of *Region ID* (see (IEEE802.16e,

(7)

of correct decoding even if all of them are received with errors.

of HARQ map can be described by the subsequent formula:

*DL*

*UL*

*N frame*

*OHSchemeIII SizeARQ \_ FB\_ III* (6)

*frame*

*N*

the following equitation:

**2.3 HARQ** 

Redundancy (IR).

probability of correct decoding.

2006)). In the simulations performed in this chapter, the *Region ID* is not considered. The actual amount of bits of *SubB* depends on the HARQ Type. Based on the (IEEE802.16e, 2006), the size of message according to the sub-bursts is following:

$$\begin{aligned} SubB\_{CC} &= \\$ + N\_{sub} \times (RCID + 20 + DIUC) \\ SubB\_{IR-CTC} &= \\$ + N\_{sub} \times (RCID + 20) \\ SubB\_{IR-CC} &= \\$ + N\_{sub} \times (R + 22 + DIUC) \end{aligned} \tag{8}$$

where *Nsub* is a number of sub-bursts; *RCID* represents a size of Reduced CID; and *DIUC* represents the size of optional field, denoted as DIUC, containing 8 bits if included.

For the case when a low number of bursts are transmitted within a frame, the utilization of so called Compact HARQ DL/UL maps enables to reduce an overhead (see (IEEE802.16e, 2006)). The overhead generated by compact version of maps is not dependent on the HARQ type. The amount of overhead can be expressed by the next equations:

$$\begin{aligned} \text{HARQOHcomp}\_{\text{DL}} &= 12 + \text{RCID} + \text{HCI} + \text{CCI} \\ \text{HARQOHcomp}\_{\text{UL}} &= 12 + \text{RCID} + \text{HCI} \end{aligned} \tag{9}$$

where *HCI* is a size of HARQ control IE (8 bits if HARQ is enabled and 4 bits if HARQ is temporary disabled); CCI is a size of CQICH control IE (16 bits if CQICH information are included and 4 bits if the information are not included).

The simple evaluation of equations for full and compact HARQ maps enables to determine which kind of maps generates minimum management overhead over the number of HARQ sub-bursts (see Fig. 2). As the results show, the compact version of maps is profitable for all numbers of sub-bursts in UL as well as for up to 12 sub-bursts in DL over all length of Reduced CID.

Fig. 2. Comparison of the overhead generated by compact and full HARQ maps for DL *(a)*  and UL *(b)*

On Efficiency of ARQ and HARQ Entities Interaction in WiMAX Networks 155

Each packet is transmitted either directly to the BS or over particular number of hops. The probability of block error between two stations is the same over all hops. Therefore, the overall BLER of all hops (between the MS and the BS) is calculated according to the

where *BLERhop* represents a BLER over each hop and *Nhops* is the number of overall hops between the MS and the BS. Note that *Nhops=n+1*, where *n* is the number of RS in the

If RSs are considered, the absolute level of transmitted overhead rises *n* times comparing to the direct communication without RSs. This is due to the fact that feedback information is

The setting of simulation parameters is depicted in Table 3. The evaluation is performed for BLER up to 10% per one hop. For higher BLER level, the channel is nearly unusable due to high error rate. Note that BLER of overall path from the MS to the BS is significantly increasing with rising number of hops (see (11)). The BLER of whole path from the MS to the

For more precise evaluation, the overhead of upper layer is also considered. The TCP

The user's data are transmitted in a number of frames transmitted from the BS to the MS. The overhead size is evaluated per all transmitted frames. A frame consists of one or several PDUs and a PDU itself contains one or several ARQ blocks. The frames are subsequently sent by the BS to the MS. A vector indicating positions of blocks with/without errors is created for each frame based on the given value of BLER. The MS responds to the BS by sending ARQ feedback message that includes selected ARQ scheme, ACK Type, and a vector of errors in the transmission. According to the feedback message, the BS retransmits erroneous blocks as soon as possible, but not before the third frame after the original transmission. The size of user's data in a DL frame is kept the same within each simulation

BS is 27% if three hops are taken into account and if BLER of a hop is 10%.

protocol is assumed for an error correction by upper layer**.** 

*Nhops BLER 1 BLER MS BS hop* (11)

Fig. 4. Link level simulation scenario

transmitted individually over each hop.

drop (1024 bytes or 4096 bytes).

following formula:

communication chain.

The relation between BLER (for ARQ confirmation) and PER (for HARQ confirmation) is defined according to (Provvedi et al., 2004) by following equation:

$$PER = 1 - \left(1 - BLER\right)^{N\_{blank}} \tag{10}$$

#### **2.4 Cooperation of ARQ with HARQ to reduce signaling overhead**

The mutual interaction consists in exchanging of information on successful packets transmission between ARQ and HARQ entities (see Fig. 3). Therefore, the data confirmed by HARQ need not to be confirmed again by ARQ process.

Fig. 3. Principle of ARQ and HARQ cooperation

At the side of transmitter, both ARQ and HARQ are implemented and applied on data. Similarly, the receiving side evaluates both ARQ and HARQ as well. However, ARQ process on the receiving side need not to transmit all requests related to the corrupted data if these data are already requested to be retransmitted by HARQ. The same way is applied for acknowledgement of data. In other words, data confirmed by HARQ are not further confirmed by ARQ. The information on ACK/NACK is delivered to HARQ processes at the side of original transmitter and HARQ just provides information of ACK/NACK data to the ARQ process at transmitter. Therefore, a part of overhead due to duplicated confirmation of data delivery is saved.

#### **3. System model and simulation parameters**

The simulator, developed in MATLAB, focuses on the evaluation of overhead generated by ARQ and HARQ procedure in the uplink direction by one user (see Fig. 4). In simulations, we assume direct communication between MS and BS as well as multihop communication using RSs.

Fig. 4. Link level simulation scenario

The relation between BLER (for ARQ confirmation) and PER (for HARQ confirmation) is

The mutual interaction consists in exchanging of information on successful packets transmission between ARQ and HARQ entities (see Fig. 3). Therefore, the data confirmed by

At the side of transmitter, both ARQ and HARQ are implemented and applied on data. Similarly, the receiving side evaluates both ARQ and HARQ as well. However, ARQ process on the receiving side need not to transmit all requests related to the corrupted data if these data are already requested to be retransmitted by HARQ. The same way is applied for acknowledgement of data. In other words, data confirmed by HARQ are not further confirmed by ARQ. The information on ACK/NACK is delivered to HARQ processes at the side of original transmitter and HARQ just provides information of ACK/NACK data to the ARQ process at transmitter. Therefore, a part of overhead due to duplicated confirmation of

The simulator, developed in MATLAB, focuses on the evaluation of overhead generated by ARQ and HARQ procedure in the uplink direction by one user (see Fig. 4). In simulations, we assume direct communication between MS and BS as well as multihop communication

*PER 1 (1 BLER)Nblocks* (10)

defined according to (Provvedi et al., 2004) by following equation:

**2.4 Cooperation of ARQ with HARQ to reduce signaling overhead** 

HARQ need not to be confirmed again by ARQ process.

Fig. 3. Principle of ARQ and HARQ cooperation

**3. System model and simulation parameters** 

data delivery is saved.

using RSs.

Each packet is transmitted either directly to the BS or over particular number of hops. The probability of block error between two stations is the same over all hops. Therefore, the overall BLER of all hops (between the MS and the BS) is calculated according to the following formula:

$$BLER\_{MS-BS} = \left(1 - BLER\_{h\eta\eta}\right)^{N\_{h\eta\eta}}\tag{11}$$

where *BLERhop* represents a BLER over each hop and *Nhops* is the number of overall hops between the MS and the BS. Note that *Nhops=n+1*, where *n* is the number of RS in the communication chain.

If RSs are considered, the absolute level of transmitted overhead rises *n* times comparing to the direct communication without RSs. This is due to the fact that feedback information is transmitted individually over each hop.

The setting of simulation parameters is depicted in Table 3. The evaluation is performed for BLER up to 10% per one hop. For higher BLER level, the channel is nearly unusable due to high error rate. Note that BLER of overall path from the MS to the BS is significantly increasing with rising number of hops (see (11)). The BLER of whole path from the MS to the BS is 27% if three hops are taken into account and if BLER of a hop is 10%.

For more precise evaluation, the overhead of upper layer is also considered. The TCP protocol is assumed for an error correction by upper layer**.** 

The user's data are transmitted in a number of frames transmitted from the BS to the MS. The overhead size is evaluated per all transmitted frames. A frame consists of one or several PDUs and a PDU itself contains one or several ARQ blocks. The frames are subsequently sent by the BS to the MS. A vector indicating positions of blocks with/without errors is created for each frame based on the given value of BLER. The MS responds to the BS by sending ARQ feedback message that includes selected ARQ scheme, ACK Type, and a vector of errors in the transmission. According to the feedback message, the BS retransmits erroneous blocks as soon as possible, but not before the third frame after the original transmission. The size of user's data in a DL frame is kept the same within each simulation drop (1024 bytes or 4096 bytes).

On Efficiency of ARQ and HARQ Entities Interaction in WiMAX Networks 157

presented also if HARQ and innovative ARQ scheme proposed in (Becvar & Bestak, 2011)

As can be observed from Fig. 5 - Fig. 16, the scenario where ARQ and HARQ interact outperforms all other scenarios. Additional minor improvement is achieved by using innovative ARQ instead of conventional ARQ. However, this improvement is noticeable only as long as ARQ\_Block\_Size is low (e.g., 16 bytes), PDU Size is higher (e.g., 16 blocks)

(a) (b) Fig. 5. ARQ & HARQ Overhead vs.BLER for ARQ\_Block\_Size = 16 B, PDU Size = 1, Size of user data = 1024 B/frame, 4 HARQ retrans., HARQ Type: CC, 1 hop *(a)* / 3 hops *(b)* 

(a) (b)

Fig. 6. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 1024 B, PDU Size = 1, Size of user data = 1024 B/frame, 4 HARQ retrans., HARQ Type: CC, 1 hop *(a)* / 3 hops *(b)* 

(in figures *ARQ PIII*) are simultaneously utilized.

and mutual interaction of ARQ and HARQ is considered.


This process is repeated until all frames are sent to the MS and the MS confirms error-free reception of all blocks. The same vectors indicating positions of blocks with/without errors are considered in all ARQ schemes.

Table 2. Simulation parameters for ARQ and HARQ

The maximum number of HARQ retransmissions is set to 2 and 4. Both types of HARQ, Chase Combining (CC) and Incremental Redundancy (IR) are considered in evaluations. The Convolutional Turbo Code (CTC) is considered in evaluation if IR HARQ is performed.

#### **4. Results**

The results are separated into several groups according to the number of hops (left-hand and right-hand figures corresponds to one and three hops respectively), HARQ Type (CC HARQ in Fig. 5 - Fig. 10 and IR-CTC in Fig. 11 - Fig. 16), and maximum number of HARQ retransmissions for higher clarity. The figures are grouped into set of six figures with the same HARQ Type, with the same maximum number of retransmissions, and further, with varying number of hops, ARQ\_Block\_Size, and PDU Size. The results are presented in form of figures showing the overhead generated due to ACK/NACK by HARQ and ARQ for 2000 continuously transmitted frames. The expressed overhead is normalized to the overhead generated by conventional IEEE802.16e ARQ (in figures noted as *Conv. ARQ*) for error free channel, using Selective ACK (in figures marked as *SACK*) together with HARQ while no interaction between both is considered. The cumulative ACK (*CACK*) is also taken into account in figures. All figures also depict results for both techniques while interaction is not enabled (without interaction - in figures denoted *w/o int.*) and while the interaction is enabled (with interaction - in figures noted as *w int.*). The overhead for the same cases is

This process is repeated until all frames are sent to the MS and the MS confirms error-free reception of all blocks. The same vectors indicating positions of blocks with/without errors

The maximum number of HARQ retransmissions is set to 2 and 4. Both types of HARQ, Chase Combining (CC) and Incremental Redundancy (IR) are considered in evaluations. The Convolutional Turbo Code (CTC) is considered in evaluation if IR HARQ is performed.

The results are separated into several groups according to the number of hops (left-hand and right-hand figures corresponds to one and three hops respectively), HARQ Type (CC HARQ in Fig. 5 - Fig. 10 and IR-CTC in Fig. 11 - Fig. 16), and maximum number of HARQ retransmissions for higher clarity. The figures are grouped into set of six figures with the same HARQ Type, with the same maximum number of retransmissions, and further, with varying number of hops, ARQ\_Block\_Size, and PDU Size. The results are presented in form of figures showing the overhead generated due to ACK/NACK by HARQ and ARQ for 2000 continuously transmitted frames. The expressed overhead is normalized to the overhead generated by conventional IEEE802.16e ARQ (in figures noted as *Conv. ARQ*) for error free channel, using Selective ACK (in figures marked as *SACK*) together with HARQ while no interaction between both is considered. The cumulative ACK (*CACK*) is also taken into account in figures. All figures also depict results for both techniques while interaction is not enabled (without interaction - in figures denoted *w/o int.*) and while the interaction is enabled (with interaction - in figures noted as *w int.*). The overhead for the same cases is

are considered in all ARQ schemes.

**Parameter Value** 

Number of frames 2000

BLER per hop [%] 0 – 10

ARQ\_Block\_Size [bytes] 16 – 1024

ARQ ACK Types Selective, Cumulative

Number of hops 1, 3

PDU size [blocks] 1 – 16

Max. HARQ retransmissions 2, 4

HARQ packet/burst size 1 PDU

RCID [bits**]** 7

Table 2. Simulation parameters for ARQ and HARQ

**4. Results** 

Size of data in each DL frame [bytes] 1024

HARQ Type CC, IR-CTC

presented also if HARQ and innovative ARQ scheme proposed in (Becvar & Bestak, 2011) (in figures *ARQ PIII*) are simultaneously utilized.

As can be observed from Fig. 5 - Fig. 16, the scenario where ARQ and HARQ interact outperforms all other scenarios. Additional minor improvement is achieved by using innovative ARQ instead of conventional ARQ. However, this improvement is noticeable only as long as ARQ\_Block\_Size is low (e.g., 16 bytes), PDU Size is higher (e.g., 16 blocks) and mutual interaction of ARQ and HARQ is considered.

Fig. 5. ARQ & HARQ Overhead vs.BLER for ARQ\_Block\_Size = 16 B, PDU Size = 1, Size of user data = 1024 B/frame, 4 HARQ retrans., HARQ Type: CC, 1 hop *(a)* / 3 hops *(b)* 

Fig. 6. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 1024 B, PDU Size = 1, Size of user data = 1024 B/frame, 4 HARQ retrans., HARQ Type: CC, 1 hop *(a)* / 3 hops *(b)* 

On Efficiency of ARQ and HARQ Entities Interaction in WiMAX Networks 159

The following group of figures represents the results of IR\_CTC HARQ for maximum four

(a) (b)

Fig. 9. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 1024 B, PDU Size = 1, Size of user data = 1024 B/frame, 2 HARQ retrans., HARQ Type: CC, 1 hop *(a)* / 3 hops *(b)* 

(a) (b)

Fig. 10. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 16 B, PDU Size = 16, Size of user data = 1024 B/frame, 2 HARQ retrans., HARQ Type: CC, 1 hop *(a)* / 3 hops *(b)* 

HARQ retransmissions.

While no interaction between HARQ and ARQ entities is enabled, the difference between conventional innovative ARQ is more significant. The reduction of overhead is more appreciable for lower number of hops or higher ARQ\_Block\_Size. The improvement achieved by innovative ARQ in comparison to scenario using conventional ARQ without interaction is due the fact that the ARQ PIII generates lower overhead while the packets are delivered without errors.

The first group of figures shows the results of CC HARQ for maximum four HARQ retransmissions.

Fig. 7. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 16 B, PDU Size = 16, Size of user data = 1024 B/frame, 4 HARQ retrans., HARQ Type: CC, 1 hop *(a)* / 3 hops *(b)*

The next group of figures depicts the results of CC HARQ for maximum two HARQ retransmissions.

Fig. 8. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 16 B, PDU Size = 1, Size of user data = 1024 B/frame, 2 HARQ retrans., HARQ Type: CC, 1 hop *(a)* / 3 hops *(b)*

While no interaction between HARQ and ARQ entities is enabled, the difference between conventional innovative ARQ is more significant. The reduction of overhead is more appreciable for lower number of hops or higher ARQ\_Block\_Size. The improvement achieved by innovative ARQ in comparison to scenario using conventional ARQ without interaction is due the fact that the ARQ PIII generates lower overhead while the packets are

The first group of figures shows the results of CC HARQ for maximum four HARQ

(a) (b) Fig. 7. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 16 B, PDU Size = 16, Size of

The next group of figures depicts the results of CC HARQ for maximum two HARQ

(a) (b) Fig. 8. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 16 B, PDU Size = 1, Size of

user data = 1024 B/frame, 2 HARQ retrans., HARQ Type: CC, 1 hop *(a)* / 3 hops *(b)*

user data = 1024 B/frame, 4 HARQ retrans., HARQ Type: CC, 1 hop *(a)* / 3 hops *(b)*

delivered without errors.

retransmissions.

retransmissions.

The following group of figures represents the results of IR\_CTC HARQ for maximum four HARQ retransmissions.

Fig. 9. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 1024 B, PDU Size = 1, Size of user data = 1024 B/frame, 2 HARQ retrans., HARQ Type: CC, 1 hop *(a)* / 3 hops *(b)* 

Fig. 10. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 16 B, PDU Size = 16, Size of user data = 1024 B/frame, 2 HARQ retrans., HARQ Type: CC, 1 hop *(a)* / 3 hops *(b)* 

On Efficiency of ARQ and HARQ Entities Interaction in WiMAX Networks 161

(a) (b)

Fig. 13. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 16 B, PDU Size = 16, Size of user data = 1024 B/frame, 4 HARQ retrans., HARQ Type: IR, 1 hop *(a)* / 3 hops *(b)*

The last group of figures represents the results of IR\_CTC HARQ for maximum two HARQ

(a) (b) Fig. 14. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 16 B, PDU Size = 1, Size of

user data = 1024 B/frame, 2 HARQ retrans., HARQ Type: IR, 1 hop *(a)* / 3 hops *(b)* 

retransmissions.

Fig. 11. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 16 B, PDU Size = 1, Size of user data = 1024 B/frame, 4 HARQ retrans., HARQ Type: IR, 1 hop *(a)* / 3 hops *(b)* 

Fig. 12. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 1024 B, PDU Size = 1, Size of user data = 1024 B/frame, 4 HARQ retrans., HARQ Type: IR, 1 hop *(a)* / 3 hops *(b)* 

(a) (b)

Fig. 11. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 16 B, PDU Size = 1, Size of

(a) (b)

Fig. 12. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 1024 B, PDU Size = 1, Size of user data = 1024 B/frame, 4 HARQ retrans., HARQ Type: IR, 1 hop *(a)* / 3 hops *(b)* 

user data = 1024 B/frame, 4 HARQ retrans., HARQ Type: IR, 1 hop *(a)* / 3 hops *(b)* 

Fig. 13. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 16 B, PDU Size = 16, Size of user data = 1024 B/frame, 4 HARQ retrans., HARQ Type: IR, 1 hop *(a)* / 3 hops *(b)*

The last group of figures represents the results of IR\_CTC HARQ for maximum two HARQ retransmissions.

Fig. 14. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 16 B, PDU Size = 1, Size of user data = 1024 B/frame, 2 HARQ retrans., HARQ Type: IR, 1 hop *(a)* / 3 hops *(b)* 

On Efficiency of ARQ and HARQ Entities Interaction in WiMAX Networks 163



The chapter investigates the efficiency of ARQ and HARQ mechanism used in WiMAX. The conventional ARQ, innovative ARQ, and HARQ are described. In addition, their cooperation is contemplated for stand alone operation of both ARQ and HARQ as well as

The results demonstrate that if the HARQ and ARQ are enabled and no mutual interaction between both entities is considered, the difference between conventional ARQ and innovative ARQ is significant. The exact level of overhead reduction depends heavily on the setting of the ARQ and HARQ parameters. The local interaction between ARQ and HARQ enables additional reduction of the overhead. If interaction is considered, the significant improvement by using innovative ARQ instead of conventional ARQ is achieved only while

This work has been performed in the framework of the FP7 project ROCKET IST-215282 STP, which is funded by the European Community. The Authors would like to acknowledge the contributions of their colleagues from ROCKET Consortium (http://www.ict-rocket.eu).

Becvar, Z. & Bestak, R. (2011). Overhead of ARQ mechanism in IEEE 802.16 networks.

Hoymann, C. (2005). Analysis and performance evaluation of the OFDM-based metropolitan

IEEE 802.16e. (2006). Air Interface for Fixed and Mobile Broadband Wireless Access

Combined Fixed and Mobile Operation in Licensed Bands. Standard IEEE IEEE 802.16j. (2009). Air Interface for Broadband Wireless Access Systems, Amendment 1:

Multihop Relay Specification. Standard IEEE

*Telecommunication Systems*, Vol.46, No.4, (March 2010), pp. 353-367, ISSN 1018-

area network IEEE 802.16. *Computer Networks*, Vol.49, No.3, pp. 341-363, ISSN 1389-

Systems: Amendment for Physical and Medium Access Control Layers for

retransmissions. Hence the ARQ overhead in similar for both cases.

reason for this conclusion is the same as explained in the previous bullet.

reduction by this parameter is only minor.

**5. Conclusions** 

for cooperation between both.

**6. Acknowledgment** 

**7. References** 

4864

1286

ARQ\_Block\_Size is low and PDU Size is high.

is delivered with errors is increasing as well; however influence of the level of overhead

Fig. 15. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 1024 B, PDU Size = 1, Size of user data = 1024 B/frame, 2 HARQ retrans., HARQ Type: IR, 1 hop *(a)* / 3 hops *(b)* 

Fig. 16. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 16 B, PDU Size = 16, Size of user data = 1024 B/frame, 2 HARQ retrans., HARQ Type: IR, 1 hop *(a)* / 3 hops *(b)*

The impact of individual parameters observed from the previous figures on the efficiency of overhead reduction can be summarized into the following concluding remars:


is delivered with errors is increasing as well; however influence of the level of overhead reduction by this parameter is only minor.


#### **5. Conclusions**

162 Advanced Transmission Techniques in WiMAX

(a) (b) Fig. 15. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 1024 B, PDU Size = 1, Size of user data = 1024 B/frame, 2 HARQ retrans., HARQ Type: IR, 1 hop *(a)* / 3 hops *(b)* 

(a) (b) Fig. 16. ARQ & HARQ Overhead vs. BLER for ARQ\_Block\_Size = 16 B, PDU Size = 16, Size of user data = 1024 B/frame, 2 HARQ retrans., HARQ Type: IR, 1 hop *(a)* / 3 hops *(b)*

The impact of individual parameters observed from the previous figures on the efficiency of




overhead reduction can be summarized into the following concluding remars:

is not influenced by a number of hops if interaction is enabled.

amount of packets not corrected by HARQ.

The chapter investigates the efficiency of ARQ and HARQ mechanism used in WiMAX. The conventional ARQ, innovative ARQ, and HARQ are described. In addition, their cooperation is contemplated for stand alone operation of both ARQ and HARQ as well as for cooperation between both.

The results demonstrate that if the HARQ and ARQ are enabled and no mutual interaction between both entities is considered, the difference between conventional ARQ and innovative ARQ is significant. The exact level of overhead reduction depends heavily on the setting of the ARQ and HARQ parameters. The local interaction between ARQ and HARQ enables additional reduction of the overhead. If interaction is considered, the significant improvement by using innovative ARQ instead of conventional ARQ is achieved only while ARQ\_Block\_Size is low and PDU Size is high.

#### **6. Acknowledgment**

This work has been performed in the framework of the FP7 project ROCKET IST-215282 STP, which is funded by the European Community. The Authors would like to acknowledge the contributions of their colleagues from ROCKET Consortium (http://www.ict-rocket.eu).

#### **7. References**


**9** 

*Ukraine* 

**Performance Analysis and** 

Oleksandr Tsopa and Vladimir Shokalo

Oleksii Strelnitskiy, Oleksandr Strelnitskiy, Oleksandra Dudka,

Currently, *WiMAX* systems acquired widespread adoption, to help organize the network at *MAN (Metropolitan access network)* level. It is assumed that in the next decade, the performance of such systems can achieve 50 bits/second/Hz*.* This is due primarily to the fact that in *MAN* network large amounts of multimedia confidential information of high quality are transmitted. So it requires not only high speeds, but also the appropriate level of noise immunity. Therefore, the performance and noise immunity are the two main indicators of the radio system quality. The radio waves propagation (*RWP*) models in radio channel play a decisive role in their calculation. Several mechanisms of radio wave propagation are known, including the so-called street-wave channels (Fabricio, 2005) (Wei, 2007), that was already noted in the development of analog communication (Porrat, 2002). These channels are identified by the authors as wavelength channels formed by architectural buildings (WCAB) (Strelnitskiy, 2007). However, the existing mathematical models are extremely complicated, so they become ineffective in the calculation of large branching networks of urban and this fact makes it difficult to assess these characteristics. Performance is evaluated as the ratio of transmission rate to channel bandwidth and noise immunity is defined as the probability of a bit (BER) or packet errors (PER). The authors propose general

approach to the WCAB representation in the form of microwave multipole.

estimates are given in the conclusion section.

**architectural buildings** 

Then the multipolar model of branched-line and outdoor radio channels is described. It allows us to calculate the attenuation of radio waves in the city streets. The reliability of the model is established by comparing the results of numerical and field experiments conducted by the authors.Performance and noise immunity of *WiMAX* communication channel

To form a mathematical model of *WCAB* we propose to adopt the following approach. Fig. 1

**2. Mathematical model of branched wavelengths, which are formed by** 

shows the city district fragment with the base station (BS) placed on the square.

**1. Introduction** 

*Kharkiv National University of Radio Electronics (KhNURE), Kharkiv,* 

**Noise Immunity** *WiMax*

**Radio Channel** 


### **Performance Analysis and Noise Immunity** *WiMax* **Radio Channel**

Oleksii Strelnitskiy, Oleksandr Strelnitskiy, Oleksandra Dudka, Oleksandr Tsopa and Vladimir Shokalo *Kharkiv National University of Radio Electronics (KhNURE), Kharkiv, Ukraine* 

#### **1. Introduction**

164 Advanced Transmission Techniques in WiMAX

Kang, M. S. & Jang, J. (2006). Performance evaluation of IEEE 802.16d ARQ algorithms with

Lee, B.G. & Choi,S. (2008). *Broadband Wireless Access and Local Networks: Mobile WiMAX and* 

Maheshwari, S., Boariu, A. & Bacciccola, A. (2008). ARQ/HARQ inter-working to reduce

Martikainen, H.; Sayenko, A.; Alanen, O. & Tykhomyrov, V. (2008). Optimal MAC PDU size

Nuaymi, L. (2007). *WiMAX: Technology for Broadband Wireless Access*. West Sussex:

Provvedi, L.; Rattray, C.; Hofmann, J. & Parolari, S. (2004). Provision of MBMS over the

Sayenko, A.; Martikainen, H. & Puchko, A. (2008). Performance comparison of HARQ and

Sengupta, S.; Chatterjee, M. & Ganguly, S. (2008). Improving Quality of VoIP Streams over

Sengupta, S.; Chatterjee, M.; Ganguly, S. & Izmailov, R. (2005). Exploiting MAC Flexibility

Tykhomyrov, V.; Sayenko, A.; Martikainen, H.; Alanen, O. & Hämäläinen, T. (2007).

*Generation Teletraffic and Wired/Wireless Advanced Networking*, pp. 148-161

WiMax. *IEEE Transactions on Computers*, Vol.57, No.2, 145-156

*Conference on 3G Mobile Communication Technologies*, pp. 494- 498, 2004. Sambale, K.; Becvar, Z. & Ulvan, A. (2008). Identification of the MAC/PHY key

Republic of Korea, 2006

*WiFi*. USA: Artech House

*on QoS in Multiservice IP Networks,* pp. 66-71

08/1142

STP

Canada

2005

Wiley&son Ltd.

NS-2 simulator, *Proceeding of Asia-Pacific Conference on Communications*. Busan,

the ARQ feedback overhead. Contribution to IEEE 802.16m No. C802.16m-

in IEEE 802.16, *Proceeding of 4th International Telecommunication Networking Workshop* 

GERAN: technical solutions and performance. *Proceeding of Fifth IEEE International* 

reconfiguration parameters, ICT ROCKET project milestone 5M2, ICT-215282

ARQ mechanisms in IEEE 802.16 networks, *Proceeding of International symposium on Modeling, analysis and simulation of wireless and mobile systems*. Vancouver,

in WiMAX for Media Streaming, *Proceedings of World of Wireless Mobile and Multimedia Networks*, pp. 338-343, Taormina - Giardini Naxos, Italy, 13-16 June

Performance Evaluation of the IEEE 802.16 ARQ Mechanism, *Proceeding ofNext* 

Currently, *WiMAX* systems acquired widespread adoption, to help organize the network at *MAN (Metropolitan access network)* level. It is assumed that in the next decade, the performance of such systems can achieve 50 bits/second/Hz*.* This is due primarily to the fact that in *MAN* network large amounts of multimedia confidential information of high quality are transmitted. So it requires not only high speeds, but also the appropriate level of noise immunity. Therefore, the performance and noise immunity are the two main indicators of the radio system quality. The radio waves propagation (*RWP*) models in radio channel play a decisive role in their calculation. Several mechanisms of radio wave propagation are known, including the so-called street-wave channels (Fabricio, 2005) (Wei, 2007), that was already noted in the development of analog communication (Porrat, 2002). These channels are identified by the authors as wavelength channels formed by architectural buildings (WCAB) (Strelnitskiy, 2007). However, the existing mathematical models are extremely complicated, so they become ineffective in the calculation of large branching networks of urban and this fact makes it difficult to assess these characteristics. Performance is evaluated as the ratio of transmission rate to channel bandwidth and noise immunity is defined as the probability of a bit (BER) or packet errors (PER). The authors propose general approach to the WCAB representation in the form of microwave multipole.

Then the multipolar model of branched-line and outdoor radio channels is described. It allows us to calculate the attenuation of radio waves in the city streets. The reliability of the model is established by comparing the results of numerical and field experiments conducted by the authors.Performance and noise immunity of *WiMAX* communication channel estimates are given in the conclusion section.

#### **2. Mathematical model of branched wavelengths, which are formed by architectural buildings**

To form a mathematical model of *WCAB* we propose to adopt the following approach. Fig. 1 shows the city district fragment with the base station (BS) placed on the square.

Performance Analysis and Noise Immunity *WiMax* Radio Channel 167

In this case, the problem is formulated as follows. Let there be a chain, equivalent to *WCAB* and containing the n external arms. It is required to determine the amplitude and phase of the field in a certain section of the circuit produced in accordance with these *WCAB* coordinates. In general, the circuit is excited with any number of arms (Fig. 2., where *ai, bi* – normalized amplitudes of the incident and reflected waves). For example in the case of Fig.

The problem is solved by the method (Gostev, 1997). Let us isolate the circuit section in which you want to determine the amplitude and phase of the signal. Conventionally, we break the transmission line at this point (Fig. 2). Let us denote additional arms through n+1

In (Gostev, 1997) it is shown that if an equivalent multipole is excited with the *i-th* arm, then the values of the normalized amplitudes of the reflected waves in the ( 1) *n* and ( 2) *n*

1, 1, 2 1, 1 2,

2, 2, 1 2, 2 1,

*ni nn nn ni n i nn nn nn nn*

*ni nn nn ni n i nn nn nn nn*

*S S SS b a S S SS* 

*S S SS b a S S SS* 

If the circuit is excited with the arms, the resultant wave can be written as (Gostev, 1997):

1, 2 2, 1 2, 2 1, 1

1, 2 2, 1 2, 2 1, 1

, (1)

. (2)

1 2 . *n n bb b* (3)

and n+2, and the matrix of the resulting multipole - via[ ]( , 1,2,..., 1, 2) *S ij n n ij* .

(1 )

(1 )

(1 )(1 )

(1 )(1 )

Fig. 2. Equivalent multipole

arms will be written as:

1 the number of excitation sources is 4.

1

2

In this case the resulting wave in the cross section

Fig. 1. Fragment of the city region

Numerals indicate segments of streets. For example, the designation 1-2 should be read: the second part of the first street. *BS* radiates waves of spherical front (*WSF*). Further we shall consider (based upon the Huygens principle) that in the radial streets (denoted by numbers 1-4 in Fig. 1) the radiated spherical wave is transformed into a series of waves with the locally flat front (*LFF*). A further approach is to use the following approximations. The street is represented by several continuous, homogeneous and smooth surfaces, which form the guide system with losses. Straight-line segment of length *l* of this guide system is replaced by an equivalent two-wire line segment. Wave resistance and wave factor of this line are equal to the characteristic impedance and wave ratio of free space. Equivalence should be determined by equality of power transferred to the real and equivalent systems, i.e. largest attenuation. The attenuation is determined by losses in *RWP*, calculated by *RWP LAN-MAN* model (Strelnitskiy, 2008), phase – by value *l* . As a result, segments of the line are easily represented by matrices of the quadripole scattering [*S*] (Gostev, 1997).

The properties of these segments are as follows: all lines have equivalent impedance equal to *Z*<sup>0</sup> because they spread a wave of T-type; street intersection (for example, 2, 3 с 5, 6 on Fig. 1) is a set of included equivalent line segments and in terms of circuit theory it is a power distribution system (*PDS*).

*СРМ* with *n* equal divisions of channels, described by the matrix [S] of ideal multipoles, together with segments of lines with losses constitute a particular scheme, the calculation of which can be performed by cyclic algorithm. As a result, the considered scheme is equivalent to a multipole (Gostev, 1997) (Fig. 2), which can be used to determine the amplitude of the field at any point of *WCAB*.

Fig. 2. Equivalent multipole

Numerals indicate segments of streets. For example, the designation 1-2 should be read: the second part of the first street. *BS* radiates waves of spherical front (*WSF*). Further we shall consider (based upon the Huygens principle) that in the radial streets (denoted by numbers 1-4 in Fig. 1) the radiated spherical wave is transformed into a series of waves with the locally flat front (*LFF*). A further approach is to use the following approximations. The street is represented by several continuous, homogeneous and smooth surfaces, which form the guide system with losses. Straight-line segment of length *l* of this guide system is replaced

equal to the characteristic impedance and wave ratio of free space. Equivalence should be determined by equality of power transferred to the real and equivalent systems, i.e. largest attenuation. The attenuation is determined by losses in *RWP*, calculated by *RWP LAN-MAN*

The properties of these segments are as follows: all lines have equivalent impedance equal to *Z*<sup>0</sup> because they spread a wave of T-type; street intersection (for example, 2, 3 с 5, 6 on Fig. 1) is a set of included equivalent line segments and in terms of circuit theory it is a

*СРМ* with *n* equal divisions of channels, described by the matrix [S] of ideal multipoles, together with segments of lines with losses constitute a particular scheme, the calculation of which can be performed by cyclic algorithm. As a result, the considered scheme is equivalent to a multipole (Gostev, 1997) (Fig. 2), which can be used to determine the

*l* . As a result, segments of the line are easily

of this line are

by an equivalent two-wire line segment. Wave resistance and wave factor

represented by matrices of the quadripole scattering [*S*] (Gostev, 1997).

Fig. 1. Fragment of the city region

model (Strelnitskiy, 2008), phase – by value

amplitude of the field at any point of *WCAB*.

power distribution system (*PDS*).

In this case, the problem is formulated as follows. Let there be a chain, equivalent to *WCAB* and containing the n external arms. It is required to determine the amplitude and phase of the field in a certain section of the circuit produced in accordance with these *WCAB* coordinates. In general, the circuit is excited with any number of arms (Fig. 2., where *ai, bi* – normalized amplitudes of the incident and reflected waves). For example in the case of Fig. 1 the number of excitation sources is 4.

The problem is solved by the method (Gostev, 1997). Let us isolate the circuit section in which you want to determine the amplitude and phase of the signal. Conventionally, we break the transmission line at this point (Fig. 2). Let us denote additional arms through n+1 and n+2, and the matrix of the resulting multipole - via[ ]( , 1,2,..., 1, 2) *S ij n n ij* .

In (Gostev, 1997) it is shown that if an equivalent multipole is excited with the *i-th* arm, then the values of the normalized amplitudes of the reflected waves in the ( 1) *n* and ( 2) *n* arms will be written as:

$$b\_{n+1} = \frac{S\_{n+1,i}(1 - S\_{n+1,n+2}) + S\_{n+1,n+1}S\_{n+2,i}}{(1 - S\_{n+1,n+2})(1 - S\_{n+2,n+1}) - S\_{n+2,n+2}S\_{n+1,n+1}} \cdot a\_{i'} \tag{1}$$

$$b\_{n+2} = \frac{S\_{n+2,i}(1 - S\_{n+2,n+1}) + S\_{n+2,n+2}S\_{n+1,i}}{(1 - S\_{n+1,n+2})(1 - S\_{n+2,n+1}) - S\_{n+2,n+2}S\_{n+1,n+1}} \cdot a\_i \,. \tag{2}$$

In this case the resulting wave in the cross section

$$b\_{\Sigma} = b\_{n+1} + b\_{n+2}.\tag{3}$$

If the circuit is excited with the arms, the resultant wave can be written as (Gostev, 1997):

Performance Analysis and Noise Immunity *WiMax* Radio Channel 169

The matrix is written on the assumption that the quadripole is consistent with the characteristic impedance of free space and reciprocal. Further let us assume that we need to determine signal strength along the street 2 (Fig. 1). To simplify the calculations, we assume that the wave processes occurring along the street 2, will be affected only by the adjacent streets 1, 3, 5 and 6. Then the electrical circuit of the equivalent multipole will have the form

It is easy to see that the diagram in Fig. 4, consists of the three basic elements: quadripole – cascade connection of attenuator and the ideal line segment - the ideal six-pole and the ideal eight-pole. We assign respectively, numbers 1, 2, 3 for the above basic elements and depict

From the above equivalent circuit, it follows that the scattering matrix of the equivalent multipole can be obtained by applying the cyclic algorithms for cascade connection of

the equivalent circuit of the equivalent multipole (Fig. 4, b).

quadripole 1 and six pole 2, and also quadripole 1 and eight-pole 3.

a) b)

12 *S S <sup>S</sup> <sup>A</sup>* ,

where (1) [ ] *ij <sup>S</sup>* , (2) [ ] *ij <sup>S</sup>* – scattering matrixes of quadripole and eight-pole.

(1) (2) 2 (2) 22 13

*A* ,

(1) (2) 2 (1) 12 11

*A* ,

*S S*( ) *S S*

Fig. 4. The electric circuit (a) and equivalent circuit (b) of *WCAB* multipole

Let's give the formula for calculating the scattering parameters of the quadripole 1 and

13 *S S <sup>S</sup> <sup>A</sup>* ,

*SSS S S*

(1) (2) 12 13

*A* , (1) (2)

(1) (2) (2) (2) 22 12 13

22 22

*S S*( ) *S S*

(1) (2) 2 (2) 22 12

*A* , (6)

22 11 *A* 1 *S S* (7)

(1) (2) 12 12

23 23

The formulae describing connection of the quadripole 1 and eight-pole 3 (Gostev, 1997)

shown in Fig. 4, a.

eight-pole 2 (Gostev, 1997)

11 11

( ) *S S S S*

33 33

$$b\_{\Sigma} = \frac{\sum\_{i=1}^{k} \left[ S\_{n+1,i} \left( 1 - S\_{n+1,n+2} + S\_{n+2,n+2} \right) + S\_{n+2,i} \left( 1 - S\_{n+2,n+1} + S\_{n+1,n+1} \right) \right]}{\left( 1 - S\_{n+1,n+2} \right) \left( 1 - S\_{n+2,n+1} \right) - S\_{n+2,n+2} S\_{n+1,n+1}} \cdot a\_i. \tag{4}$$

The *Sn*<sup>2</sup> , *Sn*<sup>1</sup> coefficients in the expressions (1, 2) are defined by cyclic algorithms given in (Gostev, 1997). For their use it is necessary to make the scheme which will be replaced by the multi-pole circuit. It is compiled on the basis of a multipole electrical circuit.

For example, in the case shown in the Fig. 1, we can make calculations from the block diagram shown in Fig. 3.

Fig. 3. *WCAB* block diagram

The scheme consists of a base station transmitter which is connected through its antenna, which has *NT* emitters, with a spatial power distributor (*SPD*). On the *SPD* outputs there are the *NR* receiving antennas, connected by the corresponding transitions to the equivalent line of *PDS* in the 1... *T TK* reference plane. Other reference planes are connected to the channel receiver, as well as loads of equivalent lines *ZL* , equal to their characteristic impedance.

Amplitudes *<sup>i</sup> a* in the 1... *T TK* reference plane depend on the relative position of the transmitting and receiving antennas and their radiation patterns. The coordinates of the receiving antennas in the cross sections of streets determine the positions of longitudinal sections of the streets along which the attenuation calculations are carried out. Let us represent the equivalent circuit for a part of urban area (Fig. 1). We assume that each segment of the street with length of *ir* may be substituted by a segment of an ideal two-wire line, that is connected with the attenuator in cascade, its damping value *<sup>i</sup>* at *RWP* is equal to the damping on the street segment with length *ir* .

The scattering matrix of quadripole equivalent to a cascading line connection and the attenuator is given by:

$$[S(r\_i)] = \begin{bmatrix} 0 & \sqrt{a\left(\frac{r\_i}{r\_0}\right)} \\ \sqrt{a\left(\frac{r\_i}{r\_0}\right)} & 0 \end{bmatrix} e^{-i\beta r\_i} \cdot \tag{5}$$

multi-pole circuit. It is compiled on the basis of a multipole electrical circuit.

1 1

1

*i*

diagram shown in Fig. 3.

Fig. 3. *WCAB* block diagram

impedance.

attenuator is given by:

*k*

. 1 1

*i*

(5)

*<sup>i</sup>* at *RWP* is equal

(4)

1, 2 2, 1 2, 2 1, 1

1, 1, 2 2, 2 2, 2, 1 1, 1

 

*nn nn nn nn*

*nj nn nn ni nn nn*

The *Sn*<sup>2</sup> , *Sn*<sup>1</sup> coefficients in the expressions (1, 2) are defined by cyclic algorithms given in (Gostev, 1997). For their use it is necessary to make the scheme which will be replaced by the

For example, in the case shown in the Fig. 1, we can make calculations from the block

The scheme consists of a base station transmitter which is connected through its antenna, which has *NT* emitters, with a spatial power distributor (*SPD*). On the *SPD* outputs there are the *NR* receiving antennas, connected by the corresponding transitions to the equivalent line of *PDS* in the 1... *T TK* reference plane. Other reference planes are connected to the channel receiver, as well as loads of equivalent lines *ZL* , equal to their characteristic

Amplitudes *<sup>i</sup> a* in the 1... *T TK* reference plane depend on the relative position of the transmitting and receiving antennas and their radiation patterns. The coordinates of the receiving antennas in the cross sections of streets determine the positions of longitudinal sections of the streets along which the attenuation calculations are carried out. Let us represent the equivalent circuit for a part of urban area (Fig. 1). We assume that each segment of the street with length of *ir* may be substituted by a segment of an ideal two-wire

The scattering matrix of quadripole equivalent to a cascading line connection and the

[ ( )] .

 

0

*i*

*S r e r r*

0

0

*i*

*r r*

*i*

*i r*

0

line, that is connected with the attenuator in cascade, its damping value

*i*

to the damping on the street segment with length *ir* .

*SSS SSS b a S S SS*

The matrix is written on the assumption that the quadripole is consistent with the characteristic impedance of free space and reciprocal. Further let us assume that we need to determine signal strength along the street 2 (Fig. 1). To simplify the calculations, we assume that the wave processes occurring along the street 2, will be affected only by the adjacent streets 1, 3, 5 and 6. Then the electrical circuit of the equivalent multipole will have the form shown in Fig. 4, a.

It is easy to see that the diagram in Fig. 4, consists of the three basic elements: quadripole – cascade connection of attenuator and the ideal line segment - the ideal six-pole and the ideal eight-pole. We assign respectively, numbers 1, 2, 3 for the above basic elements and depict the equivalent circuit of the equivalent multipole (Fig. 4, b).

From the above equivalent circuit, it follows that the scattering matrix of the equivalent multipole can be obtained by applying the cyclic algorithms for cascade connection of quadripole 1 and six pole 2, and also quadripole 1 and eight-pole 3.

Fig. 4. The electric circuit (a) and equivalent circuit (b) of *WCAB* multipole

Let's give the formula for calculating the scattering parameters of the quadripole 1 and eight-pole 2 (Gostev, 1997)

$$S\_{11} = S\_{11}^{(1)} + \frac{(S\_{12}^{(1)})^2 S\_{11}^{(2)}}{A}, \\ S\_{12} = \frac{S\_{12}^{(1)} S\_{12}^{(2)}}{A}, \\ S\_{13} = \frac{S\_{12}^{(1)} S\_{13}^{(2)}}{A}, \\ S\_{22} = S\_{22}^{(2)} + \frac{S\_{22}^{(1)} (S\_{12}^{(2)})^2}{A}, \\ \tag{6}$$

$$\mathbf{S}\_{33} = \mathbf{S}\_{33}^{(2)} + \frac{\mathbf{S}\_{22}^{(1)}(\mathbf{S}\_{13}^{(2)})^2}{A}, \ \mathbf{S}\_{23} = \mathbf{S}\_{23}^{(2)} + \frac{\mathbf{S}\_{22}^{(1)}\mathbf{S}\_{12}^{(2)}\mathbf{S}\_{13}^{(2)}}{A}, \ \mathbf{A} = \mathbf{1} - \mathbf{S}\_{22}^{(1)}\mathbf{S}\_{11}^{(2)}\tag{7}$$

where (1) [ ] *ij <sup>S</sup>* , (2) [ ] *ij <sup>S</sup>* – scattering matrixes of quadripole and eight-pole.

The formulae describing connection of the quadripole 1 and eight-pole 3 (Gostev, 1997)

Performance Analysis and Noise Immunity *WiMax* Radio Channel 171

wave propagation in urban environments. In the end, for example, modulation types are revealed which are peculiar to one or another level of signal/noise (*S/N*) ratio at the reception point. However, the experimental results described in the mentioned works are of particular nature. They cannot be used to construct a general *RWP* model of *WiMAX* wireless channels, both because of the limited number of experimental studies, and because of the lack of their systematization on any grounds. In particular, the mechanism of wave propagation along urban wave channels formed by the architectural building is not studied. Increased knowledge of the laws of propagation of *WCAB* wireless networks with *WiMAX* technology, especially in city streets, is important in connection with putting into operation mobile *WiMAX* systems at the present time and requires conducting extensive experimental

The purpose of the work in this subsection was to conduct experiments and analyze their results in the propagation of *DITS* signals with *WiMAX* technology along the street *WCAB* in a large industrial city (Kharkov). The map of the part of Kharkov where the investigations

work. Some of the experiments are done within the present study.

Fig. 5. The Map of the study area in Kharkov (BS - base station location)

were made is shown in Fig. 5.

$$\hat{S}\_{11} = S\_{11}^{\mathcal{M}} + \frac{4S\_{12}^{\mathcal{M}}S\_{21}^{\mathcal{M}}S\_{11}}{1 - S\_{11}(S\_{22}^{\mathcal{M}} + \sum\_{N=2}^{N} S\_{2,N+1})},\\ \hat{S}\_{21} = \frac{S\_{21}^{\mathcal{M}}S\_{21}^{\mathcal{M}}}{1 - S\_{11}(S\_{22}^{\mathcal{M}} + \sum\_{N=2}^{N} S\_{2,N+1})}.\tag{8}$$

Let is denote

$$\mathbf{S}\_{11}^{\mathbf{M}} = \mathbf{S}\_{11}^{\mathbf{\hat{O}}} \; \; \; \mathbf{S}\_{21}^{\mathbf{M}} = \mathbf{S}\_{21}^{\mathbf{\hat{O}}} \; \; \; \mathbf{4} \\ \mathbf{S}\_{12}^{\mathbf{M}} = \mathbf{S}\_{12}^{\mathbf{\hat{O}}} \; \; \; \; \mathbf{S}\_{22}^{\mathbf{M}} + \sum\_{N=2}^{N} \mathbf{S}\_{2,N+1} = \mathbf{S}\_{22}^{\mathbf{\hat{O}}} \; \; \; \tag{9}$$

Considering notation (9) expression (10) can be written this way

$$
\hat{S}\_{11} = S\_{11}^{\hat{\Omega}} + \frac{S\_{12}^{\hat{\Omega}} S\_{21}^{\hat{\Omega}} S\_{11}}{1 - S\_{11} S\_{22}^{\hat{\Omega}}}, \ \hat{S}\_{21} = \frac{S\_{21}^{\hat{\Omega}} S\_{21}}{1 - S\_{11} S\_{22}^{\hat{\Omega}}}.\tag{10}
$$

The scattering matrix of an equivalent quadripole of *i-*row of the schema will be:

$$
\begin{bmatrix} S^{\mathcal{D}(i)} \end{bmatrix} = \begin{bmatrix} S\_{11}^{(i)} & N\_i S\_{12}^{(i)} \\ S\_{21}^{(i)} & S\_{22}^{(i)} + \sum\_{N\_i=2}^{N\_i} S\_{2,N\_i+1} \end{bmatrix}' \tag{11}
$$

where ( ) , *i <sup>i</sup> <sup>j</sup> S* – are the scattering coefficients of the divider or quadripole of *i*-row; *N* –amount of inputs of *i-* row element.

The above formulae (1) - (11) constitute a *WCAB* mathematical model.

#### **3. Model of street wave channels formed by architectural buildings when**  *WiMAX* **system works in the city**

In this section, the general *WCAB* model developed in Section 2 is refined for the case of outdoor radio channels taking in consideration the characteristics of *WiMAX* antenna systems and *RWP* canyon model.

This section also describes the attenuation of radio waves along the street radio channels of the central district of Kharkov. The measurements were made at *3.5 GHz* with the *WiMAX* base station and a created mobile laboratory. A comprehensive analysis of the results is completed - the mechanism of formation of field distribution along the streets is elucidated. Comparative results of calculations and experiments are presented. Practical suitability of the created model in the problems of forecasting of attenuation in outdoor *WCAB* is proved.

#### **3.1 Experimental studies of attenuation in the street wave channels formed by architectural buildings when** *WiMAX* **system works in the city**

The design of digital wireless communication systems is based largely on the design of the radio channel. The accurate model of radio channel as we know from (Hata, 1980), is always based on the experiment.

For the case of digital information transmission system (*DITS*) with *WiMAX*-technology there appeared a number of articles (Fabricio, 2005); they highlight some issues of radio

21

ˆ

, M M

*S S <sup>S</sup>*

12 12 4*S S* , <sup>M</sup> <sup>Э</sup>

1 *S S <sup>S</sup>*

*N*

*N S SS* 

Э 21 21 21 Э

> , *<sup>i</sup> i*

*N*

11 22

21 21

1( ) *N*

*SS S* 

*N*

*N*

. (9)

*S S* . (10)

(11)

. (8)

M 11 22 2, 1 2

22 2, 1 22 2

*N*

*N*

Э Э Э 12 21 11 11 11 Э

1 *SSS S S*

The scattering matrix of an equivalent quadripole of *i-*row of the schema will be:

11 22

() ()

**3. Model of street wave channels formed by architectural buildings when** 

In this section, the general *WCAB* model developed in Section 2 is refined for the case of outdoor radio channels taking in consideration the characteristics of *WiMAX* antenna

This section also describes the attenuation of radio waves along the street radio channels of the central district of Kharkov. The measurements were made at *3.5 GHz* with the *WiMAX* base station and a created mobile laboratory. A comprehensive analysis of the results is completed - the mechanism of formation of field distribution along the streets is elucidated. Comparative results of calculations and experiments are presented. Practical suitability of the created model in the problems of forecasting of attenuation in outdoor *WCAB* is proved.

The design of digital wireless communication systems is based largely on the design of the radio channel. The accurate model of radio channel as we know from (Hata, 1980), is always

For the case of digital information transmission system (*DITS*) with *WiMAX*-technology there appeared a number of articles (Fabricio, 2005); they highlight some issues of radio

**3.1 Experimental studies of attenuation in the street wave channels formed by** 

**architectural buildings when** *WiMAX* **system works in the city** 

*Э i N i i*

ˆ

( ) ( ) 11 12

*i i i*

*S NS*

21 22 2, 1 2

*SS S* 

*N*

*<sup>i</sup> <sup>j</sup> S* – are the scattering coefficients of the divider or quadripole of *i*-row; *N* –amount

*i*

*S S* ,

M M

1( ) *N*

*SS S* 

*N*

M 11 22 2, 1 2

<sup>M</sup> <sup>Э</sup> *S S* 11 11 , <sup>M</sup> <sup>Э</sup> *S S* 21 21 , <sup>M</sup> <sup>Э</sup>

Considering notation (9) expression (10) can be written this way

( )

The above formulae (1) - (11) constitute a *WCAB* mathematical model.

*S*

ˆ

M 12 21 11

11 11

Let is denote

where ( ) , *i*

of inputs of *i-* row element.

*WiMAX* **system works in the city** 

systems and *RWP* canyon model.

based on the experiment.

4 ˆ

*SSS S S*

wave propagation in urban environments. In the end, for example, modulation types are revealed which are peculiar to one or another level of signal/noise (*S/N*) ratio at the reception point. However, the experimental results described in the mentioned works are of particular nature. They cannot be used to construct a general *RWP* model of *WiMAX* wireless channels, both because of the limited number of experimental studies, and because of the lack of their systematization on any grounds. In particular, the mechanism of wave propagation along urban wave channels formed by the architectural building is not studied.

Increased knowledge of the laws of propagation of *WCAB* wireless networks with *WiMAX* technology, especially in city streets, is important in connection with putting into operation mobile *WiMAX* systems at the present time and requires conducting extensive experimental work. Some of the experiments are done within the present study.

The purpose of the work in this subsection was to conduct experiments and analyze their results in the propagation of *DITS* signals with *WiMAX* technology along the street *WCAB* in a large industrial city (Kharkov). The map of the part of Kharkov where the investigations were made is shown in Fig. 5.

Fig. 5. The Map of the study area in Kharkov (BS - base station location)

Performance Analysis and Noise Immunity *WiMax* Radio Channel 173

а) b)

a) b)

Fig. 9. The measurement data on Lenin Avenue

Fig. 8. Polar patterns: а) for BS antennas; б) for subscriber station (SS) antennas

Let us describe the experiments conducted, their results are shown in Fig. 9-13. The first experiment is to measure the radiation pattern along the maximum base station (along Lenin Ave). The results of the experiments in the form of dependency of *S(r)* and *S/N(r)* are shown in Fig. 9 а, b. In the experiments, the maximum distance *r* was equal to 4 km, which corresponded to the maximum range of confident communication. In the figure dots

c)

For measurements a mobile laboratory was created, its general form is shown in Fig. 6, and its structure in Fig. 7.

Fig. 6. General view of the mobile laboratory

The mobile laboratory is equipped as follows: *WiMAX Breeze-Max 3500* (*Alvarion*) subscriber station, «Asus» notebook, NovAtel SS-11 *GPS* receiver, voltage transformer VT – 12V/220V and storage battery SB–12V. To measure the signal/noise ratio (*S/N*) and signal level (S) we used special software interface that was provided by «Alternet». *WiMAX* base station (BS) was placed at the altitude of *hБС* = 80 m. (Gasprom building, Fig. 6, right picture).

Fig. 7. The structure of the mobile laboratory

In the BS four quadrant antennas is used. One of the sectors of the polar pattern (PP) (Fig. 8, а) serves the area shown in the map (Fig. 5), the direction of maximum radiation is almost identical with the direction (orientation) of Lenin Ave.

For measurements a mobile laboratory was created, its general form is shown in Fig. 6, and

The mobile laboratory is equipped as follows: *WiMAX Breeze-Max 3500* (*Alvarion*) subscriber station, «Asus» notebook, NovAtel SS-11 *GPS* receiver, voltage transformer VT – 12V/220V and storage battery SB–12V. To measure the signal/noise ratio (*S/N*) and signal level (S) we used special software interface that was provided by «Alternet». *WiMAX* base station (BS)

In the BS four quadrant antennas is used. One of the sectors of the polar pattern (PP) (Fig. 8, а) serves the area shown in the map (Fig. 5), the direction of maximum radiation is almost

was placed at the altitude of *hБС* = 80 m. (Gasprom building, Fig. 6, right picture).

its structure in Fig. 7.

Fig. 6. General view of the mobile laboratory

Fig. 7. The structure of the mobile laboratory

identical with the direction (orientation) of Lenin Ave.

Fig. 8. Polar patterns: а) for BS antennas; б) for subscriber station (SS) antennas

Let us describe the experiments conducted, their results are shown in Fig. 9-13. The first experiment is to measure the radiation pattern along the maximum base station (along Lenin Ave). The results of the experiments in the form of dependency of *S(r)* and *S/N(r)* are shown in Fig. 9 а, b. In the experiments, the maximum distance *r* was equal to 4 km, which corresponded to the maximum range of confident communication. In the figure dots

Fig. 9. The measurement data on Lenin Avenue

Performance Analysis and Noise Immunity *WiMax* Radio Channel 175

At the same time the data received are well correlated with the results presented in (Porrat, 2002) (Fig. 10, c and Fig. 11) in changes of modulation types on the track. Analyzing them together with the data of the experiment, we can conclude that in the tested part of Kharkov the transfer of information with *WiMAX* wireless communication system is carried out with

The distances shown in Fig. 11, а-d, were measured from Н points in the direction of arrows (Fig. 5). In this case the main maximum PP of subscriber station (Fig. 8, b) was set along the street axis (approximately at the angle of *90о* to the direction of maximum BS radiation). For this reason the signal level decreased by *30 dB* compared to its level on Lenin Ave, which

Fig. 11. Approximation of the measured distributions of fields are known functions

According to *WiMAX* radio access technology on the base station sector antennas with wide PP are used (Fig. 8, а), but subscriber stations have embedded antenna with narrow PP and

This feature of the *WiMAX* apparatus allows us to offer a new method of experimental evidence for the existence of the wave channels and comparison of the signal levels *S* and signal/noise *S/N* ratio, created at the receiver due to different mechanisms of propagation.

the rate of 1-2 Mbps.

corresponds to PP value at *Θ = 90о*.

low level of back lobe reception (Fig. 8, b).

represented the data of single measurements at the reference distance *r0 = 100 m* and with *100 m* step. At each point of measurement the aperture of subscriber station (SS) antenna, fixed on a tripod (Fig. 6), was placed perpendicular to the direction of maximum reception at the height of 1.5 m above the street cover. Initial measurements were averaged and smoothed using the «*Origin 6.1*» software. The processed results as solid curves are shown in Fig. 9 а, b. The same curves are shown in conjunction in Fig. 9, c. The fig shows the pattern of change of modulation type along the route. On the distances axis segments with one or another kind of modulation (*QPSK 1/2, BPSK 3/4, and BPSK 1/2*) are shown.

It is known that the *WiMAX* equipment is adaptive and allows you to maintain a constant transmission rate (or *S/N* level) of digital stream with a decrease in the signal (Fabricio, 2005).

From the presented data it follows that when S <-65 dB the adjustment does work and the *S/N* ratio decreases with the distance at the same rate as the signal level. This result significantly refines the capabilities of *WiMAX* for adaptation, since in (Balvinder, 2006) it is shown that the lower limit of adaptation is the *-75 dB* signal level.

Fig. 10. Measuring the level of *S*, *S/N* values and modulation types outside the main lobe of PP of base station antenna

represented the data of single measurements at the reference distance *r0 = 100 m* and with *100 m* step. At each point of measurement the aperture of subscriber station (SS) antenna, fixed on a tripod (Fig. 6), was placed perpendicular to the direction of maximum reception at the height of 1.5 m above the street cover. Initial measurements were averaged and smoothed using the «*Origin 6.1*» software. The processed results as solid curves are shown in Fig. 9 а, b. The same curves are shown in conjunction in Fig. 9, c. The fig shows the pattern of change of modulation type along the route. On the distances axis segments with

It is known that the *WiMAX* equipment is adaptive and allows you to maintain a constant transmission rate (or *S/N* level) of digital stream with a decrease in the signal (Fabricio,

From the presented data it follows that when S <-65 dB the adjustment does work and the *S/N* ratio decreases with the distance at the same rate as the signal level. This result significantly refines the capabilities of *WiMAX* for adaptation, since in (Balvinder, 2006) it is

Fig. 10. Measuring the level of *S*, *S/N* values and modulation types outside the main lobe of

one or another kind of modulation (*QPSK 1/2, BPSK 3/4, and BPSK 1/2*) are shown.

shown that the lower limit of adaptation is the *-75 dB* signal level.

2005).

PP of base station antenna

At the same time the data received are well correlated with the results presented in (Porrat, 2002) (Fig. 10, c and Fig. 11) in changes of modulation types on the track. Analyzing them together with the data of the experiment, we can conclude that in the tested part of Kharkov the transfer of information with *WiMAX* wireless communication system is carried out with the rate of 1-2 Mbps.

The distances shown in Fig. 11, а-d, were measured from Н points in the direction of arrows (Fig. 5). In this case the main maximum PP of subscriber station (Fig. 8, b) was set along the street axis (approximately at the angle of *90о* to the direction of maximum BS radiation). For this reason the signal level decreased by *30 dB* compared to its level on Lenin Ave, which corresponds to PP value at *Θ = 90о*.

Fig. 11. Approximation of the measured distributions of fields are known functions

According to *WiMAX* radio access technology on the base station sector antennas with wide PP are used (Fig. 8, а), but subscriber stations have embedded antenna with narrow PP and low level of back lobe reception (Fig. 8, b).

This feature of the *WiMAX* apparatus allows us to offer a new method of experimental evidence for the existence of the wave channels and comparison of the signal levels *S* and signal/noise *S/N* ratio, created at the receiver due to different mechanisms of propagation.

Performance Analysis and Noise Immunity *WiMax* Radio Channel 177

Based on the above reasoning, we can easily conclude that in this case the *RWP* mechanism by *WCAB* acts. As before, the measurements were made at the height of the receiving antenna 1,5 *AC h м* over the street surface. Further experiments showed that the intensity of the signals at points 1 and 3 differ in *–(10÷15) dB*, i.e. contribution of diffraction mechanism to the intensity of the signal is more than an order of a magnitude smaller than

The established fact of the interference of counter propagating waves in the street channels in the presence of diffraction field component can explain the pattern of change in signal attenuation along the streets. Fig. 11 shows the measured signal and signal/noise levels for a number of streets in conjunction with the curves of decrease of *S* and *S/N* under the laws

*r r* <sup>0</sup> / or *r r* <sup>0</sup> / . It is easy to see that the experimental curves fall off more slowly than

*r r* <sup>0</sup> / (Fig. 11, а, b, c) or even than *r r* <sup>0</sup> / (Fig. 11, в). These dependencies as it is known from (Grudinskaya, 1967), are characteristic for Fresnel and Relay zones at the *RWP* over the

The reducing of the extent of decrease of the power flux density of the signal *P* in the experiment compared with the above case we explain as follows. For simplicity, we assume that the phasing of the two interfering flows in a street corridor, and the diffraction component of the field is such that the vector sum can be replaced by algebraic one. Then the expressions for the damping power of the street channel from the normalized distance k

*<sup>r</sup> PD r M l L r*

In (12) it is indicated: *PD*max – is the maximum power flux density at the point 1, when *r = r0* and the energy moves in the direction of point 2 (Fig. 12); *k = 1, 2…l*, where *l* – number of sections the street of length 0 *l r* was divided into; max max 0 *M*( ) ( )/ ( ) *l PD l PD r* , where max *PD l*( ) – maximum power flux density at point 2 when energy moves in the direction of point 1 (Fig. 14); max *L r PD r PD* ( ) ( )/ *<sup>d</sup>* , where ( ) *PD r <sup>d</sup>* – power flux density in the street due

( )*r* at *l* = 10, M = 0,1 , 0*r* 100 m are given assuming that

. (12)

*r r* <sup>0</sup> / . Depending on

( ) 1 () ( ) ( 1) *n n*

the *RWP* mechanism by *WCAB* 

of <sup>2</sup>

reflecting surface.

can be easily written as:

to diffraction of the radio channel.

polynomial of *n* power (Fig. 10,а).

In Fig. 13 calculations of value

*L(r)* = *const* = 0,1, and *n* = 2.

nondirectional antennas.

Curve  0

max

*r PD k l k*

( )*r* at the decreasing site is well described by the function 1.5

the values of *M* [0,1] and *l* the pattern of measuring the field distribution along a particular street can be described either by the inverse power function (Fig. 11,а), or by

We obtained experimental field distributions along the street *WCAB* that well match with the data in (Porrat, 2002). Here they also conducted measurement of radio waves attenuation in street channels (Ottava), only at *900 MHz* frequency and with the help of

<sup>2</sup>

Let us consider Fig. 12. Here BS is located on the longitudinal pattern of the street. Then, under the assumption that there exist wave channels, at the cross street, formed by ensembles of buildings *D1* and *D2*, two streams of energy should appear (indicated by arrows in Fig. 12). These streams are running waves moving toward each other. They interfere, forming a mixed wave.

Fig. 12. *WCAB* structure

From the above description, we get the following method of experimental proof of the existence of the wave channel. The polar pattern of the receiving antenna is sent to a maximum PP to point 1 (Fig. 12), thus recording the flow of energy moving from point 1 to point 2. Then PP maximum goes to point 2 and a reverse flow of energy is recorded. The presence of both flows indicates the existence of a wave channel. Orienting the PP maximum to point 3 (points on the walls of houses ensembles), we can detect the intensity of the signal formed by the diffraction of radio wave propagation (from BS through the roofs of the D2 ensemble of houses).

The novelty of the proposed method in comparison with known works (for example, (Volkov, 2005)) is that using the antenna with a narrow PP one can detect the direction of energy flow along the streets and separate the contributions of different *RWP* mechanisms to the received signal level.

The experimental studies by the proposed method were performed in one of the four sectors of BS.

The results of measuring the levels of signal *S* and signal/noise *S/N* are shown in Fig. 10. Here curves 1 and 3 – shows the dependence of the ratio S/N and S signal levels, respectively, along the street when moving from point 1 to point 2, and curves 2 and 4 – are the same curves, only measured when the vehicle was moving in the opposite direction (i.e. the aperture of the receiving antenna was rotated at *180º*).

Let us consider Fig. 12. Here BS is located on the longitudinal pattern of the street. Then, under the assumption that there exist wave channels, at the cross street, formed by ensembles of buildings *D1* and *D2*, two streams of energy should appear (indicated by arrows in Fig. 12). These streams are running waves moving toward each other. They

From the above description, we get the following method of experimental proof of the existence of the wave channel. The polar pattern of the receiving antenna is sent to a maximum PP to point 1 (Fig. 12), thus recording the flow of energy moving from point 1 to point 2. Then PP maximum goes to point 2 and a reverse flow of energy is recorded. The presence of both flows indicates the existence of a wave channel. Orienting the PP maximum to point 3 (points on the walls of houses ensembles), we can detect the intensity of the signal formed by the diffraction of radio wave propagation (from BS through the roofs

The novelty of the proposed method in comparison with known works (for example, (Volkov, 2005)) is that using the antenna with a narrow PP one can detect the direction of energy flow along the streets and separate the contributions of different *RWP* mechanisms

The experimental studies by the proposed method were performed in one of the four sectors

The results of measuring the levels of signal *S* and signal/noise *S/N* are shown in Fig. 10. Here curves 1 and 3 – shows the dependence of the ratio S/N and S signal levels, respectively, along the street when moving from point 1 to point 2, and curves 2 and 4 – are the same curves, only measured when the vehicle was moving in the opposite direction (i.e.

the aperture of the receiving antenna was rotated at *180º*).

interfere, forming a mixed wave.

Fig. 12. *WCAB* structure

of the D2 ensemble of houses).

to the received signal level.

of BS.

Based on the above reasoning, we can easily conclude that in this case the *RWP* mechanism by *WCAB* acts. As before, the measurements were made at the height of the receiving antenna 1,5 *AC h м* over the street surface. Further experiments showed that the intensity of the signals at points 1 and 3 differ in *–(10÷15) dB*, i.e. contribution of diffraction mechanism to the intensity of the signal is more than an order of a magnitude smaller than the *RWP* mechanism by *WCAB* 

The established fact of the interference of counter propagating waves in the street channels in the presence of diffraction field component can explain the pattern of change in signal attenuation along the streets. Fig. 11 shows the measured signal and signal/noise levels for a number of streets in conjunction with the curves of decrease of *S* and *S/N* under the laws of <sup>2</sup> *r r* <sup>0</sup> / or *r r* <sup>0</sup> / . It is easy to see that the experimental curves fall off more slowly than <sup>2</sup> *r r* <sup>0</sup> / (Fig. 11, а, b, c) or even than *r r* <sup>0</sup> / (Fig. 11, в). These dependencies as it is known from (Grudinskaya, 1967), are characteristic for Fresnel and Relay zones at the *RWP* over the reflecting surface.

The reducing of the extent of decrease of the power flux density of the signal *P* in the experiment compared with the above case we explain as follows. For simplicity, we assume that the phasing of the two interfering flows in a street corridor, and the diffraction component of the field is such that the vector sum can be replaced by algebraic one. Then the expressions for the damping power of the street channel from the normalized distance k can be easily written as:

$$\alpha \left( \frac{r\_0}{r} \right) = \frac{PD(r)}{PD\_{\text{max}}} = \left[ \frac{1}{k^n} + \frac{M(l)}{\left[ l - (k-1) \right]^n} + L(r) \right]. \tag{12}$$

In (12) it is indicated: *PD*max – is the maximum power flux density at the point 1, when *r = r0* and the energy moves in the direction of point 2 (Fig. 12); *k = 1, 2…l*, where *l* – number of sections the street of length 0 *l r* was divided into; max max 0 *M*( ) ( )/ ( ) *l PD l PD r* , where max *PD l*( ) – maximum power flux density at point 2 when energy moves in the direction of point 1 (Fig. 14); max *L r PD r PD* ( ) ( )/ *<sup>d</sup>* , where ( ) *PD r <sup>d</sup>* – power flux density in the street due to diffraction of the radio channel.

In Fig. 13 calculations of value ( )*r* at *l* = 10, M = 0,1 , 0*r* 100 m are given assuming that *L(r)* = *const* = 0,1, and *n* = 2.

Curve ( )*r* at the decreasing site is well described by the function 1.5 *r r* <sup>0</sup> / . Depending on the values of *M* [0,1] and *l* the pattern of measuring the field distribution along a particular street can be described either by the inverse power function (Fig. 11,а), or by polynomial of *n* power (Fig. 10,а).

We obtained experimental field distributions along the street *WCAB* that well match with the data in (Porrat, 2002). Here they also conducted measurement of radio waves attenuation in street channels (Ottava), only at *900 MHz* frequency and with the help of nondirectional antennas.

Performance Analysis and Noise Immunity *WiMax* Radio Channel 179

In the case of street branched radio channels general *WCAB* model, created in section 3.1, must be supplemented by the calculated damping ratios for *RWP* taking into consideration characteristics of *WiMAX* antenna system. Thus, the damping on the straight segment of the

(/) / ( )( ) ,[ ] *<sup>i</sup>*

*D D* , – amendments that allow change of *DAF* in a given direction, calculated

*PD r r rr D D D D dB*

where *D1* – maximum directional antenna factor (*DAF*) of base station (*14 dB*), *D2* – maximum *DAF* of client adapter antenna (*16,5 dB*); 0 max *PD r r PD* ( / )/ *<sup>i</sup>* – relative power flux density, which was calculated using the models described in *RWP LAN-MAN* (Strelnitskiy,

Antenna parameters significantly affect the nature and level of the signal and noise. The complexity of the problems of determining the signal strength and signal/noise level is that you need to know not only the maximum *DAF Dmax*, but also *DAF* in a particular direction

The applied methods of reducing antenna extraneous emission, and side lobe suppression leads to the complication of the analytical description of the antenna as a whole (Wheeler, 1947). In addition, such description often presents a trade secret of manufacturing companies and antenna technical data contain only its simplified polar pattern and basic characteristics. You should also take into account the fact that under real conditions of installation (the roof of a building, an antenna mast), due to the influence of the earth's surface and surrounding objects, the shape of a real PP is different from that calculated. Therefore, considering a large amount of analyzed data based on the information provided by the manufacturer of antennas, in the polar coordinate system we made polar patterns of the base station antenna (Fig. 14, a) and PP of the customer *WiMAX* antennas

 – *D(,)*.

 <sup>0</sup> 0 1122

max

, (13)

*PD*

Street name Segment length, m Width, *m*

Lenin Ave 4000 50-55 Bakulina 600 25-30 Danilevskii 1000 30-35 Lenin 1200 30-35 Lyapunov 550 30-50 Samokisha 200 25-50 Culture 600 30 Galana 500 30-40

Table 1. *WCAB* lenght and width

street should be calculated by the formula

from the mentioned considerations.

2008); 1 2 

in azimuth

(Fig. 14, b).

*i*

and the corner of the place

Fig. 13. Calculation results for formula (8)

Another proof of the validity of the results of the experiment is the data of repeated measurements shows in Fig. 11 d. These experiments were conducted one week after the first experiments. The qualitative nature of the curves is identical in both cases.

Thus, for *WCAB* of different frequencies of microwave range interference is inherent leading to the formation of mixed waves. The mathematical description of these waves is well developed in the theory of microwave circuits which is recommended for the calculation of street *WCAB* without diffractive component of the field. Another proof of the feasibility of the approach investigated to the *WCAB* analysis is presented in (Waganov, 1982).

#### **3.2 Theoretical research of attenuations in the street wave channels formed by architectural buildings at the example of** *WiMAX* **system**

When applying the *WCAB* model worked out in section 2, in the case of functioning of *WiMAX* systems, we must know the architectural features of the area where the measurements were done.

For the analysis area (Fig. 5) the following data were obtained. Height above the sea level for most of the analysis area varies smoothly from 135 to 145 m, which allows us to characterize the underlying surface as slightly undulating. The area is characterized by high building density, which makes it possible to approximate the lateral surface of wavelengths to a solid wall. Studies have shown that the material of the walls of most buildings in this case is brick. All the streets have asphalt as the underlying surface. Electrical parameters of brick can vary greatly enough and, according to the paper (Volkov, 2005), for this frequency range they are: specific permittivity *ε1=2..15*, conductivity *σ=0,002..0,01*. The conductivity of asphalt is in the same range as for bricks (*σ2=0,002..0,01*), and permittivity *ε2=2..5.* Data as for the *WCAB* length and width are given in Table. 1.


Table 1. *WCAB* lenght and width

178 Advanced Transmission Techniques in WiMAX

Another proof of the validity of the results of the experiment is the data of repeated measurements shows in Fig. 11 d. These experiments were conducted one week after the

Thus, for *WCAB* of different frequencies of microwave range interference is inherent leading to the formation of mixed waves. The mathematical description of these waves is well developed in the theory of microwave circuits which is recommended for the calculation of street *WCAB* without diffractive component of the field. Another proof of the feasibility of

When applying the *WCAB* model worked out in section 2, in the case of functioning of *WiMAX* systems, we must know the architectural features of the area where the

For the analysis area (Fig. 5) the following data were obtained. Height above the sea level for most of the analysis area varies smoothly from 135 to 145 m, which allows us to characterize the underlying surface as slightly undulating. The area is characterized by high building density, which makes it possible to approximate the lateral surface of wavelengths to a solid wall. Studies have shown that the material of the walls of most buildings in this case is brick. All the streets have asphalt as the underlying surface. Electrical parameters of brick can vary greatly enough and, according to the paper (Volkov, 2005), for this frequency range they are: specific permittivity *ε1=2..15*, conductivity *σ=0,002..0,01*. The conductivity of asphalt is in the same range as for bricks (*σ2=0,002..0,01*), and permittivity *ε2=2..5.* Data as for

first experiments. The qualitative nature of the curves is identical in both cases.

the approach investigated to the *WCAB* analysis is presented in (Waganov, 1982).

**architectural buildings at the example of** *WiMAX* **system** 

the *WCAB* length and width are given in Table. 1.

**3.2 Theoretical research of attenuations in the street wave channels formed by** 

Fig. 13. Calculation results for formula (8)

measurements were done.

In the case of street branched radio channels general *WCAB* model, created in section 3.1, must be supplemented by the calculated damping ratios for *RWP* taking into consideration characteristics of *WiMAX* antenna system. Thus, the damping on the straight segment of the street should be calculated by the formula

$$\alpha \left( r\_i \;/\; r\_0 \right) = \left( D\_1 - \Delta D\_1 \right) + \left( D\_2 - \Delta D\_2 \right) + \frac{PD(r\_i \;/\; r\_0)}{PD\_{\text{max}}}, \; \left[ dB \right], \tag{13}$$

where *D1* – maximum directional antenna factor (*DAF*) of base station (*14 dB*), *D2* – maximum *DAF* of client adapter antenna (*16,5 dB*); 0 max *PD r r PD* ( / )/ *<sup>i</sup>* – relative power flux density, which was calculated using the models described in *RWP LAN-MAN* (Strelnitskiy, 2008); 1 2 *D D* , – amendments that allow change of *DAF* in a given direction, calculated from the mentioned considerations.

Antenna parameters significantly affect the nature and level of the signal and noise. The complexity of the problems of determining the signal strength and signal/noise level is that you need to know not only the maximum *DAF Dmax*, but also *DAF* in a particular direction in azimuth and the corner of the place – *D(,)*.

The applied methods of reducing antenna extraneous emission, and side lobe suppression leads to the complication of the analytical description of the antenna as a whole (Wheeler, 1947). In addition, such description often presents a trade secret of manufacturing companies and antenna technical data contain only its simplified polar pattern and basic characteristics. You should also take into account the fact that under real conditions of installation (the roof of a building, an antenna mast), due to the influence of the earth's surface and surrounding objects, the shape of a real PP is different from that calculated. Therefore, considering a large amount of analyzed data based on the information provided by the manufacturer of antennas, in the polar coordinate system we made polar patterns of the base station antenna (Fig. 14, a) and PP of the customer *WiMAX* antennas (Fig. 14, b).

Performance Analysis and Noise Immunity *WiMax* Radio Channel 181

elevation angles of maximum radiation and reception of the transmitting and receiving antennas, respectively; *ir* – distance between the transmitter and receiver; , *H H T R* – height

*TR* are obtained from the geometrical problem in Fig.15:

 

*TR*

*D D* 

,

2

,

*DAF* relative to field strength can be calculated by the formula:

can be calculated by the formulas:


2

*b*

sin sin (sin sin cos cos cos( )) arccos , cos sin(arccos(sin sin cos cos cos( ))) *R TT R T R R Т*

 

 

*H H arctg r*

The correction that takes into account *DAF* changes in a given direction of PP, can be found from:

2 2 *DD D* ,

 

where τ – value, taking into account the reduction in antenna gain in direction Δθ,Δ

*T R T R TR RT*

*θθθ*

 

2

*b b*

*T R T R TR RT*

The width of base station antenna PP: in the horizontal plane – 90 , in the vertical plane – 8 . The width of customer adapter antenna PP: in the horizontal plane – 20 , in the vertical

for different antenna types have the form:

2 2 2 2 4 cos , 4 1 cos 1

1 1 cos , 0 65 , 90 90 , <sup>2</sup> 1 2 cos 1

,

.

 

*T T R T R RT*

. *R T*

*RT* it is necessary to change indexes *T* to *R* and *R* to *T* in (14).

 

– coefficients of directional antennas in both horizontal and vertical

*i*

 

> 

 

> 

(15)

, 20 lg τ (*dB*), (17)

(18)

(19)

(20)

   

> 

*T R* –

(14)

(16)

radiation and reception of the transmitting and receiving antennas, respectively; ,

of transmitting and receiving antennas, respectively.

 Corners

*TR* and 

*TR*

*RT* and 

and *D*

compared to the maximum gain.

To determine

where *D*

Corners Δθ,Δ

plane – 20 .

Formulas for calculation

planes, respectively.

Fig. 14. PP of the base station (а) and customer *WiMAX* antenna (b)

Mutual arrangement and orientation of antennas in the calculation of communication systems can be quite different. Therefore, the actual antenna gain at the base station in the direction of the client adapter interacting with it is defined by the angles that define the direction from the antenna of the transmitter to the receiver antenna, and vice versa, – in horizontal and vertical planes *TR* , *RT* , *TR* , *RT* respectively.

In Fig 15,a relative position of antenna polar patterns in the horizontal plane is given and the same thing is given in a plane passing perpendicular to the plane ,0, through the studied antennas in Fig. 15,b.

Fig. 15. Parameters of relative positions of antennas in the horizontal (а) and in the vertical (b) planes

In figures the following geographic coordinates are marked: *T T* , – latitude and longitude of the location of the transmitter antenna, *R R* , – latitude and longitude of the location of the receiver antenna; *T T* , – width of the transmitter PP in the horizontal and vertical planes respectively; *R R* , – the same for PP of the receiver; , *Т <sup>R</sup>* – azimuth of maximum

Mutual arrangement and orientation of antennas in the calculation of communication systems can be quite different. Therefore, the actual antenna gain at the base station in the direction of the client adapter interacting with it is defined by the angles that define the direction from the antenna of the transmitter to the receiver antenna, and vice versa, – in

In Fig 15,a relative position of antenna polar patterns in the horizontal plane is given and the

Fig. 15. Parameters of relative positions of antennas in the horizontal (а) and in the vertical

*R R* , 

– the same for PP of the receiver; ,

*T T* , 

– width of the transmitter PP in the horizontal and vertical

*Т* 

– latitude and longitude of the location of

*RT* respectively.

  through the

– latitude and longitude

*<sup>R</sup>* – azimuth of maximum

 а) b) Fig. 14. PP of the base station (а) and customer *WiMAX* antenna (b)

> *TR* , *RT* , *TR* ,

а) b)

In figures the following geographic coordinates are marked:

of the location of the transmitter antenna,

 *R R* , 

*T T* , 

same thing is given in a plane passing perpendicular to the plane ,0,

horizontal and vertical planes

studied antennas in Fig. 15,b.

(b) planes

the receiver antenna;

planes respectively;

radiation and reception of the transmitting and receiving antennas, respectively; , *T R* – elevation angles of maximum radiation and reception of the transmitting and receiving antennas, respectively; *ir* – distance between the transmitter and receiver; , *H H T R* – height of transmitting and receiving antennas, respectively.

Corners *TR* and *TR* are obtained from the geometrical problem in Fig.15:

$$\theta\_{\rm TR} = \arccos \frac{\sin \lambda\_{\rm R} - \sin \lambda\_{\rm T} (\sin \lambda\_{\rm T} \cdot \sin \lambda\_{\rm R} + \cos \lambda\_{\rm T} \cdot \cos \lambda\_{\rm R} \cdot \cos (\underline{\xi}\_{\rm R} - \underline{\xi}\_{\rm T}))}{\cos \lambda\_{\rm T} \cdot \sin (\arccos(\sin \lambda\_{\rm T} \cdot \sin \lambda\_{\rm R} + \cos \lambda\_{\rm T} \cdot \cos \lambda\_{\rm R} \cdot \cos(\underline{\xi}\_{\rm R} - \underline{\xi}\_{\rm T})))},\tag{14}$$

$$
\rho\_{\rm TR} = \operatorname{arcctg} \frac{H\_R - H\_T}{r\_i}.\tag{15}
$$

To determine *RT* and *RT* it is necessary to change indexes *T* to *R* and *R* to *T* in (14).

The correction that takes into account *DAF* changes in a given direction of PP, can be found from:

$$
\Delta D = \sqrt{D\_{A\theta}^2 + D\_{A\theta}^2} \tag{16}
$$

where *D* and *D* – coefficients of directional antennas in both horizontal and vertical planes, respectively.

*DAF* relative to field strength can be calculated by the formula:

$$D\_{A\theta\prime}D\_{A\phi} = \mathfrak{D} \cdot \lg \left(\mathfrak{r}\right) \text{ ( $dB$ )}\,,\tag{17}$$

where τ – value, taking into account the reduction in antenna gain in direction Δθ,Δ compared to the maximum gain.

Corners Δθ,Δcan be calculated by the formulas:

$$\begin{aligned} \Delta \theta\_{T,R} &= \left| \theta\_{T(R)} - \theta\_{TR(RT)} \right| \prime \\ \Delta \phi\_{T,R} &= \left| \phi\_{T(R)} - \phi\_{TR(RT)} \right| \cdot \end{aligned} \tag{18}$$

The width of base station antenna PP: in the horizontal plane – 90 , in the vertical plane – 8 . The width of customer adapter antenna PP: in the horizontal plane – 20 , in the vertical plane – 20 .

Formulas for calculation for different antenna types have the form:


$$\tau = \frac{4b^2 \cdot \cos^2 \chi}{\left(4b^2 - 1\right)\cos^2 \chi + 1},\tag{19}$$

$$b^2 = \frac{1}{2} \cdot \frac{1 - \cos^2 a}{1 - \left(\sqrt{2} \cos a - 1\right)^2}, \qquad 0^\circ \le a \le 65^\circ, \quad -90^\circ \le \chi \le 90^\circ,\tag{20}$$

Performance Analysis and Noise Immunity *WiMax* Radio Channel 183

Fig. 17. The scheme for calculating the attenuation along Lenin Avenue (*m=1* – Lenin Ave,

For signal level *S* assessment in *ir* points on Lenin Ave. the attenuator АТ(1,1) was

0 max *PD r r PD* ( / )/ *<sup>i</sup>* in formula (9) is calculated using ratio (10) for a power flux density and

1 2 120 *E*

where *Р* – emitting power at the point of transfer; *Z*0 – impedance of free space; *D*<sup>2</sup> – directional antenna factor; *R*В, *R*Г – reflection coefficients for vertical and horizontal polarizations. Similarly, attenuation for all other attenuators АТ(2,n) and АТ(3,n) were calculated. At the same time power meter in the lines that imitate the streets 2 and 3, were excluded and were attributed to the attenuator loss distribution over the entire length of the segments. Comparative data on the results of the calculation according to the method described and experimental results are shown in Fig. 18, 20, 22. Fig. 19, 21 shows a connection diagram of the streets used to calculate the attenuation along the streets in the

max

0 2 <sup>1</sup> ( ) <sup>1</sup>

*E r Re Re Re e*

*PD*

Fig. 18. Attenuation along the Lenin Ave (1 – experiment, 2 – calculation)

2

*ВГГ*

, (23)

*r r* , calculated by the formula (9). Summand

*kri kri kri kri*

. (22)

*m=2* – Romain Rolland street, *m=3* – Galana street)

the expression (11) for the field strength at the receiver.

2

*Microwave Office* application package.

*PZ D*

*r*

attributed with the decay 0 (/) *<sup>i</sup>*

where – corner ,,, *TRT R* depending on the particular antenna (transmitting - *Т*, receiving - *R*) and plane (horizontal or vertical ); – PP width in horizontal (*T R* , ) or vertical ( , *T R* ) planes;


$$\tau = \frac{(1-a)\cos\gamma + \sqrt{\left(1-a\right)^2 \cdot \cos^2\gamma + 4a}}{2},\tag{21}$$

where 0 1; *а* at 0 *а* 90 90 ; *<sup>а</sup>* 1 180 180 .

Corner is calculated by the formula: *T R i H H arctg r* Fig. 16.

Here are some examples of calculations on the proposed model by means of *Microwave Office* application package.

Fig. 16. For the calculation of the angle 

The example of a streets connection scheme used to calculate the attenuation along Lenin Avenue in the *Microwave Office* application package, is shown in Fig. 17 ( *<sup>i</sup> l r* – line segment length *ir* ; *ATmn* ( ,) – attenuators (m – street number, n – attenuator serial number for the street with m number); PI – power indicator). The scheme is activated by three sources of locally plane waves from the streets: Romain Rolland, Galana and Lenin Ave.

182 Advanced Transmission Techniques in WiMAX

 ); 

 <sup>2</sup> <sup>2</sup> 1 cos 1 cos 4 , <sup>2</sup> *a aa*

*<sup>а</sup>* 1 180 180

Here are some examples of calculations on the proposed model by means of *Microwave* 

.

*H H arctg r*

 

*i*

 

(21)

Fig. 16.

or vertical

locally plane waves from the streets: Romain Rolland, Galana and Lenin Ave.

The example of a streets connection scheme used to calculate the attenuation along Lenin Avenue in the *Microwave Office* application package, is shown in Fig. 17 ( *<sup>i</sup> l r* – line segment length *ir* ; *ATmn* ( ,) – attenuators (m – street number, n – attenuator serial number for the street with m number); PI – power indicator). The scheme is activated by three sources of

depending on the particular antenna (transmitting - *Т*,

– PP width in horizontal (

*T R* , ) or

where 

Corner 

vertical ( , *T R* 

 – corner ,,, 

receiving - *R*) and plane (horizontal

) planes;


where 0 1; *а* at 0 *а* 90 90 ;

Fig. 16. For the calculation of the angle

*Office* application package.

 

*TRT R*

is calculated by the formula: *T R*

Fig. 17. The scheme for calculating the attenuation along Lenin Avenue (*m=1* – Lenin Ave, *m=2* – Romain Rolland street, *m=3* – Galana street)

For signal level *S* assessment in *ir* points on Lenin Ave. the attenuator АТ(1,1) was attributed with the decay 0 (/) *<sup>i</sup> r r* , calculated by the formula (9). Summand 0 max *PD r r PD* ( / )/ *<sup>i</sup>* in formula (9) is calculated using ratio (10) for a power flux density and the expression (11) for the field strength at the receiver.

$$PD\_{\text{max}} = \frac{1}{2} \cdot \frac{\left| E \right|^2}{120\pi} \,\text{.}\tag{22}$$

$$\dot{E}(r) = \frac{\sqrt{P \cdot Z\_0 \cdot D\_2}}{2\pi} \cdot \left[ \frac{1}{r} \cdot \left( 1 + R\_B \cdot e^{-k \cdot r \cdot i} + R\_I \cdot e^{-k \cdot r \cdot i} + R\_{I^-} \cdot e^{-k \cdot r \cdot i} \right) \cdot e^{-k \cdot r \cdot i} \right],\tag{23}$$

where *Р* – emitting power at the point of transfer; *Z*0 – impedance of free space; *D*<sup>2</sup> – directional antenna factor; *R*В, *R*Г – reflection coefficients for vertical and horizontal polarizations. Similarly, attenuation for all other attenuators АТ(2,n) and АТ(3,n) were calculated. At the same time power meter in the lines that imitate the streets 2 and 3, were excluded and were attributed to the attenuator loss distribution over the entire length of the segments. Comparative data on the results of the calculation according to the method described and experimental results are shown in Fig. 18, 20, 22. Fig. 19, 21 shows a connection diagram of the streets used to calculate the attenuation along the streets in the *Microwave Office* application package.

Fig. 18. Attenuation along the Lenin Ave (1 – experiment, 2 – calculation)

Performance Analysis and Noise Immunity *WiMax* Radio Channel 185

Fig. 21. The scheme for calculating the attenuation along the Danilevskii street (m=1 – Lenin Ave., m=2 – Romain Rolland street, m=3 – Yaroslav Galan street, m=4 – Danilevskii street)

Fig. 22. Attenuation along the Danilevskii street (1 – experiment, 2 – calculation)

It is easily seen that the theoretical curves agree well with experimental data.

Fig. 19. The scheme for calculating the attenuation along Lenin street (m=1 – Lenin Ave., street (m=1 – Lenin Ave., m=2 – Romain Rolland street, m=3 – Yaroslav Galan street, m=4 – Lenin street, *m=5* – Novgorod street, *m=6* – Culture street, *m=7* – Baculina street, *m=8* – Ak. Lyapunov street)

Fig. 20. Attenuation along Lenin street (1 – experiment, 2 – calculation)

Fig. 19. The scheme for calculating the attenuation along Lenin street (m=1 – Lenin Ave., street (m=1 – Lenin Ave., m=2 – Romain Rolland street, m=3 – Yaroslav Galan street, m=4 – Lenin street, *m=5* – Novgorod street, *m=6* – Culture street, *m=7* – Baculina street, *m=8* – Ak.

Fig. 20. Attenuation along Lenin street (1 – experiment, 2 – calculation)

Lyapunov street)

Fig. 21. The scheme for calculating the attenuation along the Danilevskii street (m=1 – Lenin Ave., m=2 – Romain Rolland street, m=3 – Yaroslav Galan street, m=4 – Danilevskii street)

Fig. 22. Attenuation along the Danilevskii street (1 – experiment, 2 – calculation) It is easily seen that the theoretical curves agree well with experimental data.

Performance Analysis and Noise Immunity *WiMax* Radio Channel 187

Let us plot the graphs of the packet errors probability versus interference level for networks

Analyzing the results in Fig. 25 with the recommendations of the video transmission standards, we conclude that the presence of noise value less than / 0,4 *P P J S* in branched *WiMAX* channels, multimedia data can be transferred with high quality at a distance of

The model examined in this chapter can be used not only to assess the performance and noise immunity of *WiMAX* radio channel on *WCAB* conditions, but also to forecast its

*3 km,* that is very important in the construction of telemedicine networks in big cities.

а) b)

Fig. 24. Dependency of /*S N* to distance (а) and transmission rate (b) for Lenin Ave.

Fig. 25. The dependence of the probability of packet errors on the distance for different

values of noise for Lenin Ave. ( *PJ* – interference power, *PS* – signal power)

such as *MAN* (Fig. 25).

physical level security [Strelnitskiy, 2011].

#### **4. Assessment of performance and noise immunity of the** *WiMAX* **system in the city**

Let us estimate the performance and security of *WiMAX* channel in the particular example of its operation in the telemedicine system (Fig. 23). It is expedient to consider this example because in constructing the telemedicine system branched channels of street *WCAB* are used.

In this case, the base station and special ambulance are supplied with *WiMAX* equipment. Medical team transfers the data about the patient via *Wi-Fi*. Information comes from the base station to the telemedicine center and providing consultations to the medical team.

Fig. 23. Configuration of telemedicine network based on *Wi-Fi* and *WiMAX* technologies

Let us define the system efficiency, assuming that the telemedicine system serves population on Lenin Ave. The experimental and theoretical /*S N* values are shown in Fig. 5.14 (curves 1 and 2). From the comparison of these two curves it follows that the proposed model can be applied to calculate the performance of branched *WCAB* of *MAN* level (Strelnitskiy, 2009).

The rate of information transmission in *WiMAX* channel can change significantly (Fig. 24) depending on their bandwidth, which varies according to (IEEE Standard, 2004) from *1,7MHz* (curve 1) to *3,5 7MHz* (curve 2).

Comparing our results with the standards, we conclude that in the absence of interference in branched *WiMAX* channels, multimedia data can be transferred with high quality, which is very important during, for example, medical operations (Strelnitskiy, 2008).

**4. Assessment of performance and noise immunity of the** *WiMAX* **system in** 

Let us estimate the performance and security of *WiMAX* channel in the particular example of its operation in the telemedicine system (Fig. 23). It is expedient to consider this example because in constructing the telemedicine system branched channels of street *WCAB* are

In this case, the base station and special ambulance are supplied with *WiMAX* equipment. Medical team transfers the data about the patient via *Wi-Fi*. Information comes from the base station to the telemedicine center and providing consultations to the medical team.

Fig. 23. Configuration of telemedicine network based on *Wi-Fi* and *WiMAX* technologies

*1,7MHz* (curve 1) to *3,5 7MHz* (curve 2).

Let us define the system efficiency, assuming that the telemedicine system serves population on Lenin Ave. The experimental and theoretical /*S N* values are shown in Fig. 5.14 (curves 1 and 2). From the comparison of these two curves it follows that the proposed model can be applied to calculate the performance of branched *WCAB* of *MAN* level (Strelnitskiy, 2009). The rate of information transmission in *WiMAX* channel can change significantly (Fig. 24) depending on their bandwidth, which varies according to (IEEE Standard, 2004) from

Comparing our results with the standards, we conclude that in the absence of interference in branched *WiMAX* channels, multimedia data can be transferred with high quality, which is

very important during, for example, medical operations (Strelnitskiy, 2008).

**the city** 

used.

Let us plot the graphs of the packet errors probability versus interference level for networks such as *MAN* (Fig. 25).

Analyzing the results in Fig. 25 with the recommendations of the video transmission standards, we conclude that the presence of noise value less than / 0,4 *P P J S* in branched *WiMAX* channels, multimedia data can be transferred with high quality at a distance of *3 km,* that is very important in the construction of telemedicine networks in big cities.

The model examined in this chapter can be used not only to assess the performance and noise immunity of *WiMAX* radio channel on *WCAB* conditions, but also to forecast its physical level security [Strelnitskiy, 2011].

Fig. 24. Dependency of /*S N* to distance (а) and transmission rate (b) for Lenin Ave.

Fig. 25. The dependence of the probability of packet errors on the distance for different values of noise for Lenin Ave. ( *PJ* – interference power, *PS* – signal power)

Performance Analysis and Noise Immunity *WiMax* Radio Channel 189

These studies were supported by the grant from the State Foundation for Fundamental

Balvinder, B.; Eline, R. J., and Franca-Neto, L. M. (2006), RF System and Circuit Challenges

Fabricio, L. F., and Cardieri, P., (2005), Coverage Prediction and Performance Evaluation of

Gostev, V.I., Konin, V.V., and Matsepura, А.L., (1997), Linear multichannel microwave

Hata, M., (1980), Empirical formula for propagation loss in land mobile radio service, *IЕЕЕ Transactions on Vehicular Technology*, Vol. 29, No. 3, (Aug 80), pp. 317-325. IEEE Standard, (2004), Standard for Local and metropolitan area networks. Part 16: Air

Porrat, D. (2002). PhD Thesis: Radio Propagation in Hallways and Streets for UHF

Strelnitskiy, O.O.; Strelnitskiy, O.Е.; Tsopa, O.I., and Shokalo, V.М., (2011), Prediction Model

Strelnitskiy, O.O.; Tsopa, O.I. and Shokalo, V.M., (2009), Approximate Model for Estimation

Strelnitskiy, O.Е.; Tsopa, O.O.; Tsopa, O.I., and Shokalo, V.М. (2008), The variant of quality

Strelnitskiy, А.А., Strelnitskiy, А.Е., Tsopa, O.I., and Shokalo, V.М., (2008), Version of the

Strelnytskiy, A.A.; Strelnytskiy, A.E.; Tsopa, O.I., and Shokalo, V.M., (2007), The Model of

(February 2008), pp. 388-389, ISBN: 978-966-553-678-9

devices, Radio amator Publishing house, Kiev: 315 p. (in Russian)

for *WiMAX*. *Intel® Technology Journal*, Vol.08, Issue 03, (August 2004). pp. 189-201,

Wireless Metropolitan Area Networks based on IEEE 802.16. *Journal* of communication and information systems, Vol.20. No.3, (2005), pp. 132-140, ISSN:

Interface for Fixed Broadband Wireless Access Systems. IEEE P802.16-REVd/D5-

of Energy Security for the Systems of Subscriber Radio Access with Branched Street and Corridor Communications Channels. *International journal «Radioelectronics and Communications Systems»*, Allerton Press, Inc., Springer, Vol. 54, No. 2, (2011), pp.

of Efficiency and Noise Immunity of Branched Street and Corridor *Wi-Fi* and *WiMAX* Communication Channels. *International journal «Telecommunication and Radio Engineering*», Begell House, Vol. 68(17), (2009), pp. 1511-1528, ISSN: 0040-2508

increasing of video information transmission via *WIMAX* fixed connection radio channel. *Proceeding of IX International Conf. Modern problems of Radio Engineering*, *Telecommunications and Computer Science* /*TCSET'2008*/, Lviv-Slavsko, Ukraine,

model of the wideband signal attenuation in radio link when calculating the local communication networks protection, *Scientific Journal Information Protection*, Vol.

the Multiterminal Network for Attenuation Calculation of the Radio Waves in the Wave Channels of the Architectural Buildings (KNURE - WCAB Model). *Microwave & Telecommunication Technology, CriMiCo2007*. 17th International Crimean Conference Publication, (September 2007), pp. 213-214, ISBN: 978-966-335-012-7

Research, Ministry of Science and Education of Ukraine (#F25/217-2008).

**6. Acknowledgment** 

ISSN: 1535-864X

1980–6604

2004, 915 p.

Communications.

61-68, ISSN: 0735-2727.

3(39), pp. 38-43 (in Russian).

**7. References** 

#### **5. Conclusion**


high quality can be transferred at a distance of 3 km, which is very important in the construction of telemedicine networks in major cities.

#### **6. Acknowledgment**

These studies were supported by the grant from the State Foundation for Fundamental Research, Ministry of Science and Education of Ukraine (#F25/217-2008).

#### **7. References**

188 Advanced Transmission Techniques in WiMAX

1. The approach is developed to create a simplified *WCAB* model, which is based on the idea of representation branched wavelengths in the form of two-wire lines segments connected to one multipole and equivalent in the level of transmission power of street wavelengths. We propose a model to calculate the S-parameters of such multipolebased on using known cyclic algorithms and on registered losses equal to the *WRP*

2. Version of a mobile laboratory on the basis of *WiMAX* subscriber station was created and the *WRP* patterns of street *WCAB* were measured. From the comparison of the data we obtained in the signal/noise ratio with the data of *WiMAX* standard we discovered that the speed of information transmission by the *WiMAX* system operating in the city

3. New data on the possibilities of adapting the *WiMAX* system to maintain constant transmission speeds were obtained. It is shown that the adaptation of this system is realized if the signal level is above -65 dB (the lower limit declared before was -75 dB

4. Using the property of high antenna directivity of *WiMAX* client adapter we offer a new method of detection the street wavelengths and reception of the diffraction component of the field. Using this methodology, we proved the dominant existence of *WCAB* in the central area of the city of Kharkov and showed that the level of the diffraction component does not exceed - 10 dB at the chosen measurement conditions

5. It is shown that the patterns of distribution of the field along the city streets are largely predetermined by the interference of waves. A verbal description of this process is given and by comparing with the results (Porrat, 2002) a conclusion is made that it is valid for the cases of measurements in different cities and different frequencies of

6. It is concluded that *WRP* dependencies along the street channels identified in the analysis of experimental results are characteristic of microwave circuits, which suggests

8. The formula was derived for the calculation of the channel performance and the functional dependence of attenuation on the track was found, which allow to determine the performance of communication systems not worse than previously known mathematical models, but with much less time-consuming. Good agreement between calculated and experimental data gives the right to recommend the above model to calculate the performance and speed of information transmission in wireless access

9. The assessment of communication system noise immunity on *MAN* level has been given by the wave propagation in the channels formed by buildings. Comparing these results with the standards of video transmission, it was concluded that the presence of noise of less than / 0,4 *P P J S* in branched *WiMAX* channels multimedia data with high quality can be transferred at a distance of 3 km, which is very important in the

the possibility of using their well-developed theory to create a *WCAB* model. 7. The ability to predict the attenuation in branched outdoor radio channels using the

( 80 , 1,5 *БС AC h м h м* and low-rise building of analysis area).

**5. Conclusion** 

attenuation along *WCAB*.

(Balvinder, 2006).

microwave range.

systems of *MAN* level.

proposed *WCAB* model was proved.

construction of telemedicine networks in major cities.

is less than 2 Mbps at a distance of 4 km.


**10** 

*Malaysia* 

**On PAPR Reduction Techniques** 

The mobile Worldwide Interoperability for Microwave Access (Mobile WiMAX) air interface adopts orthogonal frequency division multiple access (OFDMA) as multiple access technique for its uplink (UL) and downlink (DL) to improve the multipath performance. All OFDMA based networks including mobile WiMAX experience the problem of high peak-toaverage power ratio (PAPR). The literature is replete with a large number of PAPR reduction techniques. Among them, schemes like constellation shaping, phase optimization, nonlinear companding transforms, tone reservation (TR) and tone injection (TI), clipping and filtering, partial transmit sequence (PTS), precoding based techniques, selective mapping (SLM), precoding based selective mapping (PSLM) and phase modulation transform are popular. The precoding based techniques, however, show great promise as they are simple linear techniques to implement without the need of any complex optimizations. This chapter reviews these PAPR reduction techniques and presents a Zadoff-Chu matrix transform (ZCMT) based precoding technique for PAPR reduction in mobile WiMAX systems. The mobile WiMAX systems employing random-interleaved OFDMA uplink system has been used for determining the improvement in PAPR performance of the technique. It has been further used in selective mapping (SLM) based ZCMT precoded random-interleaved OFDMA uplink system. PAPR of these systems are analyzed with the root-raised-cosine (RRC) pulse shaping to keep out-of-band radiation low and to meet the transmission spectrum mask requirement. Simulation results show that the proposed systems have low PAPR than the Walsh-Hadamard transform (WHT) precoded random-interleaved OFDMA uplink systems and the conventional random-interleaved OFDMA uplink systems. The symbol-error-rate (SER) performance of these uplink systems is also better than the conventional random-interleaved OFDMA uplink systems and at par with WHT based random-interleaved OFDMA uplink systems. The good improvement in PAPR offered by the presented systems significantly reduces the cost and the complexity of

This chapter is organized as follows: Section 2 describes the background of the randominterleaved OFDMA uplink systems and SLM based random-interleaved OFDMA uplink systems, while in section 3, we present a detailed literature review. In section 4, we present our proposed system models with improved PAPR, section 5 presents the computer

simulation results and section 6 concludes the chapter.

**1. Introduction** 

the transmitter.

**in Mobile WiMAX** 

Imran Baig and Varun Jeoti *Universiti Teknologi PETRONAS,* 


### **On PAPR Reduction Techniques in Mobile WiMAX**

Imran Baig and Varun Jeoti *Universiti Teknologi PETRONAS, Malaysia* 

#### **1. Introduction**

190 Advanced Transmission Techniques in WiMAX

Volkov, L.N., Nemirovsky, M.S., and Shinakov Y.S., (2005), Digital communication systems: basic methods and characteristics, Eco-Trends, Moscow: 392 p. (in Russian) Waganov, R.B., and Katseneleybaum, B.Z. (1982), Fundamentals of diffraction theory.

Wei, Z., (2007), Capacity analysis for multi-hop *WiMAX* relay. *Proceeding of the International* 

High school, Moscow: 244 p. (in Russian)

University of Technology Sydney, Sydney, pp. 1-4

Science, Moscow: 272 p. (in Russian) Grudinskaya, T.P., (1967), Wave propagation,

*Conference on Wireless Broadband and Ultra Wideband Communication*, (Mar. 2007),

The mobile Worldwide Interoperability for Microwave Access (Mobile WiMAX) air interface adopts orthogonal frequency division multiple access (OFDMA) as multiple access technique for its uplink (UL) and downlink (DL) to improve the multipath performance. All OFDMA based networks including mobile WiMAX experience the problem of high peak-toaverage power ratio (PAPR). The literature is replete with a large number of PAPR reduction techniques. Among them, schemes like constellation shaping, phase optimization, nonlinear companding transforms, tone reservation (TR) and tone injection (TI), clipping and filtering, partial transmit sequence (PTS), precoding based techniques, selective mapping (SLM), precoding based selective mapping (PSLM) and phase modulation transform are popular. The precoding based techniques, however, show great promise as they are simple linear techniques to implement without the need of any complex optimizations. This chapter reviews these PAPR reduction techniques and presents a Zadoff-Chu matrix transform (ZCMT) based precoding technique for PAPR reduction in mobile WiMAX systems. The mobile WiMAX systems employing random-interleaved OFDMA uplink system has been used for determining the improvement in PAPR performance of the technique. It has been further used in selective mapping (SLM) based ZCMT precoded random-interleaved OFDMA uplink system. PAPR of these systems are analyzed with the root-raised-cosine (RRC) pulse shaping to keep out-of-band radiation low and to meet the transmission spectrum mask requirement. Simulation results show that the proposed systems have low PAPR than the Walsh-Hadamard transform (WHT) precoded random-interleaved OFDMA uplink systems and the conventional random-interleaved OFDMA uplink systems. The symbol-error-rate (SER) performance of these uplink systems is also better than the conventional random-interleaved OFDMA uplink systems and at par with WHT based random-interleaved OFDMA uplink systems. The good improvement in PAPR offered by the presented systems significantly reduces the cost and the complexity of the transmitter.

This chapter is organized as follows: Section 2 describes the background of the randominterleaved OFDMA uplink systems and SLM based random-interleaved OFDMA uplink systems, while in section 3, we present a detailed literature review. In section 4, we present our proposed system models with improved PAPR, section 5 presents the computer simulation results and section 6 concludes the chapter.

On PAPR Reduction Techniques in Mobile WiMAX 193

OFDMA is widely adopted in the various communication standards like WiMAX, mobile broadband wireless access (MBWA), evolved UMTS terrestrial radio access (E-UTRA) and ultra mobile broadband (UMB). OFDMA is also a strong candidate for the wireless regional

However, OFDMA has some drawbacks, among others; the peak-to-average power ratio (PAPR) is still one of the major drawbacks in the transmitted OFDMA signal (Wang & Chen, 2004). Therefore, for zero distortion of the OFDMA signal, the high-power-amplifier (HPA) must not only operate in its linear region but also with sufficient back-off. Thus, HPA with a large dynamic range is required for OFDMA systems. These amplifiers are very expensive

Thus, if we reduce the PAPR it not only means that we are reducing the cost of OFDMA systems and reducing the complexity of analog-to-digital (A/D) and digital-to-analog (D/A) converters, but also increasing the transmit power, thus, for same range improving received signal-noise-ratio (SNR), or for the same SNR improving range. Fig.4 illustrates the block diagram of the OFDMA uplink systems. In OFDMA uplink systems the baseband modulated symbols are passed through serial-to-parallel (S/P) converter which generates

Then the subcarrier mapping of these constellations symbols can be done on in one of the subcarrier mapping mode: interleaved mode, random-interleaved mode or in localized mode respectively. After the subcarrier mapping, we get frequency domain samples: { *<sup>l</sup>* 0,1,2, 1 *Y : l ...,N } <sup>ˆ</sup>* . Mathematically, the subcarrier mapping in interleaved mode can be

where *N* : System subcarriers, *M* : User subcarriers, *Q* : Subchannels/Users, (*Q*=*N*/*M*), 0 ≤ *l* ≤ *N*-1 and *N*=*Q.M*. The subcarrier mapping in random-interleaved mode can be done

where, 0 ≤ *l* ≤ *N*-1 and *N*=*Q.M,* and *ˆ 0QQ* . The subcarrier mapping in localized mode

(1)

(2)

(3)

area networks (WRAN) and the long term evaluation advanced (LTE-Advanced).

complex vector of size *M*. We can write the complex vector of size *M* as follows:-

and are major cost component of the OFDMA systems.

Fig. 3. Localized OFDMA

done as follows:-

mathematically as follows:-

can be done mathematically as follows:-

#### **2. Background**

The mobile Worldwide Interoperability for Microwave Access (Mobile WiMAX) is a broadband wireless solution that enables the convergence of mobile and fixed broadband networks through a common wide area radio-access (RA) technology and flexible network architecture. Since January 2007, the IEEE 802.16 Working Group (WG) has been developing a new amendment of the IEEE 802.16 standard i.e. IEEE 802.16m as an advanced air interface to meet the requirements of ITU-R/IMT-Advanced for 4G systems. The mobile WiMAX air interface adopts orthogonal frequency division multiple access (OFDMA) as multiple access technique for its uplink (UL) and downlink (DL) to improve the multipath performance. The scalable OFDMA (SOFDMA) is introduced in the IEEE 802.16e amendment to support scalable channel bandwidth.

OFDMA is a multiple access version of the orthogonal frequency division multiplexing (OFDM) systems. OFDMA system splits the high speed data stream into a number of parallel low data rate streams and these low rates data streams are transmitted simultaneously over a number of orthogonal subcarriers. The key difference between OFDM and OFDMA is that instead of being allocated all of the available subcarriers, the base station assigns a subset of carriers to each user in order to accommodate several transmissions at the same time. An inherent gain of the OFDMA based systems is its ability to exploit the multiuser diversity through subchannel allocation. Additionally, OFDMA has the advantage of simple decoding at the receiver side due to the absence of inter-carrierinterference (ICI). Other benefits of OFDMA include better granularity and improved link budget in the uplink communications (Knopp & Humblet, 1995; Tse, 1997).

There are two different approaches to do subcarrier mapping in OFDMA systems, localized subcarrier mapping and distributed subcarrier mapping. The distributed subcarrier mapping can be further divided in to two modes, interleaved mode and random interleaved mode. Fig.1 shows the subcarrier mapping in interleaved mode, where the subcarriers are mapped equidistant to each other's. Fig.2 explains the subcarrier mapping in randominterleaved mode, where the subcarriers are mapped randomly based on some permutation algorithm to each other's. Fig.3 further explains the concept of localized subcarrier mapping, where the subcarrier mapping is done in adjacent.

Fig. 2. Random-Interleaved OFDMA

Fig. 3. Localized OFDMA

The mobile Worldwide Interoperability for Microwave Access (Mobile WiMAX) is a broadband wireless solution that enables the convergence of mobile and fixed broadband networks through a common wide area radio-access (RA) technology and flexible network architecture. Since January 2007, the IEEE 802.16 Working Group (WG) has been developing a new amendment of the IEEE 802.16 standard i.e. IEEE 802.16m as an advanced air interface to meet the requirements of ITU-R/IMT-Advanced for 4G systems. The mobile WiMAX air interface adopts orthogonal frequency division multiple access (OFDMA) as multiple access technique for its uplink (UL) and downlink (DL) to improve the multipath performance. The scalable OFDMA (SOFDMA) is introduced in the IEEE 802.16e

OFDMA is a multiple access version of the orthogonal frequency division multiplexing (OFDM) systems. OFDMA system splits the high speed data stream into a number of parallel low data rate streams and these low rates data streams are transmitted simultaneously over a number of orthogonal subcarriers. The key difference between OFDM and OFDMA is that instead of being allocated all of the available subcarriers, the base station assigns a subset of carriers to each user in order to accommodate several transmissions at the same time. An inherent gain of the OFDMA based systems is its ability to exploit the multiuser diversity through subchannel allocation. Additionally, OFDMA has the advantage of simple decoding at the receiver side due to the absence of inter-carrierinterference (ICI). Other benefits of OFDMA include better granularity and improved link

There are two different approaches to do subcarrier mapping in OFDMA systems, localized subcarrier mapping and distributed subcarrier mapping. The distributed subcarrier mapping can be further divided in to two modes, interleaved mode and random interleaved mode. Fig.1 shows the subcarrier mapping in interleaved mode, where the subcarriers are mapped equidistant to each other's. Fig.2 explains the subcarrier mapping in randominterleaved mode, where the subcarriers are mapped randomly based on some permutation algorithm to each other's. Fig.3 further explains the concept of localized subcarrier mapping,

budget in the uplink communications (Knopp & Humblet, 1995; Tse, 1997).

**2. Background** 

amendment to support scalable channel bandwidth.

where the subcarrier mapping is done in adjacent.

Fig. 1. Interleaved OFDMA

Fig. 2. Random-Interleaved OFDMA

OFDMA is widely adopted in the various communication standards like WiMAX, mobile broadband wireless access (MBWA), evolved UMTS terrestrial radio access (E-UTRA) and ultra mobile broadband (UMB). OFDMA is also a strong candidate for the wireless regional area networks (WRAN) and the long term evaluation advanced (LTE-Advanced).

However, OFDMA has some drawbacks, among others; the peak-to-average power ratio (PAPR) is still one of the major drawbacks in the transmitted OFDMA signal (Wang & Chen, 2004). Therefore, for zero distortion of the OFDMA signal, the high-power-amplifier (HPA) must not only operate in its linear region but also with sufficient back-off. Thus, HPA with a large dynamic range is required for OFDMA systems. These amplifiers are very expensive and are major cost component of the OFDMA systems.

Thus, if we reduce the PAPR it not only means that we are reducing the cost of OFDMA systems and reducing the complexity of analog-to-digital (A/D) and digital-to-analog (D/A) converters, but also increasing the transmit power, thus, for same range improving received signal-noise-ratio (SNR), or for the same SNR improving range. Fig.4 illustrates the block diagram of the OFDMA uplink systems. In OFDMA uplink systems the baseband modulated symbols are passed through serial-to-parallel (S/P) converter which generates complex vector of size *M*. We can write the complex vector of size *M* as follows:-

$$X = \|X\_5, X\_1, X\_2, \dots, X\_{M-1}\|^\top \tag{1}$$

Then the subcarrier mapping of these constellations symbols can be done on in one of the subcarrier mapping mode: interleaved mode, random-interleaved mode or in localized mode respectively. After the subcarrier mapping, we get frequency domain samples: { *<sup>l</sup>* 0,1,2, 1 *Y : l ...,N } <sup>ˆ</sup>* . Mathematically, the subcarrier mapping in interleaved mode can be done as follows:-

$$\hat{\mathcal{P}}\_l = \begin{cases} X\_l & \text{, } l = \mathcal{Q}, k & \mathbf{0} \le k \le M - 1 \\ \mathbf{0} & \text{otherwise} \end{cases} \tag{2}$$

where *N* : System subcarriers, *M* : User subcarriers, *Q* : Subchannels/Users, (*Q*=*N*/*M*), 0 ≤ *l* ≤ *N*-1 and *N*=*Q.M*. The subcarrier mapping in random-interleaved mode can be done mathematically as follows:-

$$\mathcal{Q}\_l = \begin{cases} X\_l & \text{, } l = \hat{Q}, k & \mathbf{0} \le k \le \mathcal{M} - 1 \\ \mathbf{0} & \text{otherwise} \end{cases} \tag{3}$$

where, 0 ≤ *l* ≤ *N*-1 and *N*=*Q.M,* and *ˆ 0QQ* . The subcarrier mapping in localized mode can be done mathematically as follows:-

On PAPR Reduction Techniques in Mobile WiMAX 195

A large number of PAPR reduction techniques have been proposed in the literature. Among them, schemes like phase optimization (Nikookar & Lidsheim, 2002), constellation shaping (Kou et al., 2007), selective mapping (SLM) (Lim et al., 2005), nonlinear companding transforms (Jiang et al., 2006), tone reservation (TR), tone injection (TI) (Mourelo., 1999; Yoo et al., 2006), partial transmit sequence (PTS) (Han & Lee, 2004; Müller & Huber, 1997; Cimini & Sollenberger, 2000; Tellambura, 2001), clipping and filtering (Wang & Tellambura, 2005; Li & Cimini, 1998; Nee & Wild, 1998), precoding based techniques (Slimane, 2007; Min & Jeoti, 2007; Baig & Jeoti, 2010a, 2010b, 2010c), precoding based selective mapping (PSLM) techniques (Baig & Jeoti, 2010a, 2010b) and phase modulation transform (Tasi et al., 2006; Thompson et al., 2008) are popular. The precoding based techniques, however, show great promise as they are simple linear techniques to implement without the need of any side

information. Additionally, the precoding based techniques take advantage of

frequency variations of the communication channel and offers substantial performance gain in fading multipath channels. In the following sub-section we focus more closely on the

The clipping techniques are simpler and commonly used to reduce the PAPR (Wang & Tellambura, 2005; Li & Cimini, 1998; Nee & Wild, 1998). These techniques apply clipping or

Fig. 4. Random-Interleaved OFDMA uplink system

PAPR reduction techniques for multicarrier transmission.

**3.1 Clipping and filtering techniques** 

**3. Literature review** 

*<sup>M</sup> T ( ).T <sup>N</sup>*

$$\mathcal{P}\_l = \begin{cases} X\_l & 0 \le l \le M - 1 \\ 0 & M \le l \le N - 1 \end{cases} \tag{4}$$

The *kth* subcarrier of each group is assigned to the *kth* user with index set {(*k*), (*Q*+*k*), …, ((*M*-1) *Q*+*k*)}. Suppose the *kth* user is assigned to subchannel *k* then the complex baseband ZCMT precoded interleaved OFDMA uplink signal for *kth* user can be written as follows:-

$$\alpha\_n^{(k)} = \sum\_{l=0}^{L-1} \hat{Y}\_l^{(k)}.e^{j2\pi \frac{(lQ+k)}{N}n}, \ n = \ 0, 1...N-1\tag{5}$$

The *kth* subcarrier of each group is assigned to the *kth* user with index set: *q ,1 q ,2 q,M 1 {( ),(Q ),...,( M 1)Q )}* , where *q ,1 q ,2 q ,M 1 {( ),( ),...,( )}* are independent and identically distributed random variables with uniform distribution on (*q*=0,1,2,…,*Q*-1). Suppose the *kth* user is assigned to sub-channel *k* then the complex baseband randominterleaved OFDMA signal for *kth* user with *N* system subcarriers and *M* user subcarriers can be written as follows:-

$$\chi\_n^{(k)} = \begin{array}{c} \sum\_{l=0}^{L-1} \binom{\hat{Y}\_l^{(k)} \dots e^{j2\pi \frac{(lQ+\nu\_{q,k})}{N}}}{N} \end{array}, n = 0, 1 \dots N - 1 \tag{6}$$

The subchannel *k* is composed of subcarriers with index set {(*kL*), (*kL*+1), (*kL*+2)… (*kL*+*L*-1)}, where *k*=0,1,2,…, *Q*-1. Suppose the *kth* user is assigned to subchannel *k* then the complex baseband ZCMT precoded localized OFDMA uplink signal for *kth* user can be written as follows:-

$$\chi\_n^{(k)} \equiv \frac{1}{\sqrt{N}} \sum\_{l=0}^{L-1} (\hat{Y}\_l^{(k)}, e^{j2\pi \frac{(kL+l)}{N}n}), \ n = 0, 1 \ldots N-1 \tag{7}$$

*(k) <sup>l</sup> Yˆ* is modulated signal on subcarrier *l* for *k*th user.

The complex passband signal of OFDMA uplink systems after the RRC pulse shaping can be written as follows:-

$$\mathbf{x}(t) = e^{j\omega\_{\varepsilon}t} \sum\_{n=0}^{N-1} \mathbf{x}\_{n}^{(k)}.\\r(t - n\vec{T})\tag{8}$$

where, *ωc* is carrier frequency, *r(t)* is baseband pulse, *<sup>M</sup> T ( ).T <sup>N</sup>* is compressed symbol duration after IFFT and *T* is symbol duration is seconds. The RRC pulse shaping filter can be defined as follows:-

$$r(t) = \frac{\sin\left(\frac{\pi t}{\overline{\tau}}(1-a)\right) + 4a\frac{t}{\overline{\tau}}\cos\left(\frac{\pi t}{\overline{\tau}}(1+a)\right)}{\frac{\pi t}{\overline{\tau}}\left(1 - \frac{16a^2t^2}{\overline{\tau}^2}\right)}\tag{9}$$

0 ≤α ≤ 1, where α is rolloff factor. The PAPR of OFDMA uplink signal in (8) with RRC pulse shaping can be written as follows:-

$$PAPR = \frac{\frac{\max\_{\mathbf{x}}}{\mathbf{x} \cdot \mathbf{t} \cdot \mathbf{s} \cdot \mathbf{T} \mathbf{t}} \left| \boldsymbol{\omega}(\mathbf{t}) \right|^{2}}{\frac{1}{N\mathbf{T}} \int\_{0}^{N\mathbf{T}} \left| \boldsymbol{\omega}(\mathbf{t}) \right|^{2} dt} \tag{10}$$

The *kth* subcarrier of each group is assigned to the *kth* user with index set {(*k*), (*Q*+*k*), …, ((*M*-1) *Q*+*k*)}. Suppose the *kth* user is assigned to subchannel *k* then the complex baseband ZCMT

The *kth* subcarrier of each group is assigned to the *kth* user with index set: *q ,1 q ,2 q,M 1 {( ),(Q ),...,( M 1)Q )}* , where *q ,1 q ,2 q ,M 1 {( ),( ),...,( )}* are independent and identically distributed random variables with uniform distribution on (*q*=0,1,2,…,*Q*-1). Suppose the *kth* user is assigned to sub-channel *k* then the complex baseband randominterleaved OFDMA signal for *kth* user with *N* system subcarriers and *M* user subcarriers can

The subchannel *k* is composed of subcarriers with index set {(*kL*), (*kL*+1), (*kL*+2)… (*kL*+*L*-1)}, where *k*=0,1,2,…, *Q*-1. Suppose the *kth* user is assigned to subchannel *k* then the complex baseband ZCMT precoded localized OFDMA uplink signal for *kth* user can be written as

The complex passband signal of OFDMA uplink systems after the RRC pulse shaping can be

duration after IFFT and *T* is symbol duration is seconds. The RRC pulse shaping filter can be

0 ≤α ≤ 1, where α is rolloff factor. The PAPR of OFDMA uplink signal in (8) with RRC pulse

precoded interleaved OFDMA uplink signal for *kth* user can be written as follows:-

be written as follows:-

*<sup>l</sup> Yˆ* is modulated signal on subcarrier *l* for *k*th user.

where, *ωc* is carrier frequency, *r(t)* is baseband pulse, *<sup>M</sup> T ( ).T <sup>N</sup>*

follows:-

*(k)*

written as follows:-

defined as follows:-

shaping can be written as follows:-

(4)

(5)

(6)

(7)

(8)

(9)

(10)

is compressed symbol

Fig. 4. Random-Interleaved OFDMA uplink system

#### **3. Literature review**

A large number of PAPR reduction techniques have been proposed in the literature. Among them, schemes like phase optimization (Nikookar & Lidsheim, 2002), constellation shaping (Kou et al., 2007), selective mapping (SLM) (Lim et al., 2005), nonlinear companding transforms (Jiang et al., 2006), tone reservation (TR), tone injection (TI) (Mourelo., 1999; Yoo et al., 2006), partial transmit sequence (PTS) (Han & Lee, 2004; Müller & Huber, 1997; Cimini & Sollenberger, 2000; Tellambura, 2001), clipping and filtering (Wang & Tellambura, 2005; Li & Cimini, 1998; Nee & Wild, 1998), precoding based techniques (Slimane, 2007; Min & Jeoti, 2007; Baig & Jeoti, 2010a, 2010b, 2010c), precoding based selective mapping (PSLM) techniques (Baig & Jeoti, 2010a, 2010b) and phase modulation transform (Tasi et al., 2006; Thompson et al., 2008) are popular. The precoding based techniques, however, show great promise as they are simple linear techniques to implement without the need of any side *<sup>M</sup> T ( ).T <sup>N</sup>* information. Additionally, the precoding based techniques take advantage of

frequency variations of the communication channel and offers substantial performance gain in fading multipath channels. In the following sub-section we focus more closely on the PAPR reduction techniques for multicarrier transmission.

#### **3.1 Clipping and filtering techniques**

The clipping techniques are simpler and commonly used to reduce the PAPR (Wang & Tellambura, 2005; Li & Cimini, 1998; Nee & Wild, 1998). These techniques apply clipping or

On PAPR Reduction Techniques in Mobile WiMAX 197

The SLM is one of the most popular PAPR reduction techniques in the literature (Lim et al., 2005). This technique is based on the phase rotations. In SLM based OFDM (SLM-OFDM) systems, a set of *V* different data blocks are created at the transmitter representing the identical information and a data block with minimum PAPR is selected for the transmission. Fig.6 shows the general block diagram of the SLM-OFDM system. Every data block is multiplied with the *V* dissimilar phase sequences, each of length *N*, *B*(*<sup>v</sup>*) = [*bv,0, bv,1, …, bv,N-1*]*T*, *v= 1, 2…V,* which results in the changed data blocks. Now suppose the altered data block for the *v*th phase sequence is given by *X*(*<sup>v</sup>*) = [*X0bv,0, X1bv,1*,…, *XN-1bv,N-1*]*T, v=1, 2… V.* Each

where*, v = 1, 2… V.* Amongst all the tailored data blocks: *x*(*<sup>v</sup>*), *v = 1, 2… V,* the data block with minimum PAPR is selected for the transmission. Side information about the selected phase sequence must be communicated to the receiver which performs the reverse operation

Fig. 6. Block diagram of OFDM system with Selective Mapping ( Han & Lee, 2005)

(14)

(15)

**3.2 Selective Mapping (SLM)** 

*<sup>v</sup> Xn* can be defined as follows:-

in order to recover the actual data block.

After applying SLM to *X*, the OFDM signal becomes as follows:-

nonlinear saturation around the peaks to lower the high PAPR produced by the multicarrier transmitter. It is straightforward to clip the signal parts that are outside the tolerable area. Clipping techniques introduces in-band or out-of-band distortions that can destroy the orthogonality between the subcarriers. Generally, the clipping operation is carried out at the transmitter. On the other hand, the receiver requires to estimate the clipping that has been carried out at the transmitter and to compensate the received OFDM symbol accordingly. Normally, most of the time no more than one clipping happens per OFDM symbol. Hence, the receiver has to approximate the size and the location of the clip. However, it is hard to get related information. After the clipping operation, the filtering operation can noticeably decrease the out-of-band radiation.

Unfortunately, the in-band distortion cannot be reduced by the filtering operation. On the other hand, the clipping can introduce some peak re-growth. So, after the clipping operation and filtering operation the signal may exceed the clipping level at some points. To decrease the peak re-growth, a repeated clipping operation and filtering operation can be carried out to obtain a desirable PAPR at the expense of computational complexity increase. Fig.5 shows represent the clipped edition of the *xp*[*m*], which can be written as follows:-

$$\chi\_{\varepsilon}^{p}[m] = \begin{cases} -\mathcal{A} & \chi^{p}[m] \le -A\\ \chi^{p}[m] & |\chi^{p}[m]| < A\\ \mathcal{A} & \chi^{p}[m] \ge A \end{cases} \tag{11}$$

or

$$
\chi\_c^p[m] = \begin{cases}
\chi^p[m] \cdot \chi^p[m] & \le -A \\
\frac{\chi^p[m]}{\varkappa^p[m]} \text{A} & \text{Otherwise}
\end{cases}
\tag{12}
$$

where *A* is pre-determined clipping level. The equation (12) can be used for both baseband complex-valued signals and passband real-valued signals and the equation (11) can only be used for the passband signals only. The clipping-ratio (CR) normalized by the root-meansquare (RMS) value σ of OFDM signal can be written as follows:-

$$\mathbf{CR} = \mathbf{A}/\sigma\tag{13}$$

Fig. 5. Block Diagram of OFDM System with Clipping and Filtering ( Cho et al., 2011)

#### **3.2 Selective Mapping (SLM)**

196 Advanced Transmission Techniques in WiMAX

nonlinear saturation around the peaks to lower the high PAPR produced by the multicarrier transmitter. It is straightforward to clip the signal parts that are outside the tolerable area. Clipping techniques introduces in-band or out-of-band distortions that can destroy the orthogonality between the subcarriers. Generally, the clipping operation is carried out at the transmitter. On the other hand, the receiver requires to estimate the clipping that has been carried out at the transmitter and to compensate the received OFDM symbol accordingly. Normally, most of the time no more than one clipping happens per OFDM symbol. Hence, the receiver has to approximate the size and the location of the clip. However, it is hard to get related information. After the clipping operation, the filtering operation can noticeably

Unfortunately, the in-band distortion cannot be reduced by the filtering operation. On the other hand, the clipping can introduce some peak re-growth. So, after the clipping operation and filtering operation the signal may exceed the clipping level at some points. To decrease the peak re-growth, a repeated clipping operation and filtering operation can be carried out to obtain a desirable PAPR at the expense of computational complexity increase. Fig.5 shows

where *A* is pre-determined clipping level. The equation (12) can be used for both baseband complex-valued signals and passband real-valued signals and the equation (11) can only be used for the passband signals only. The clipping-ratio (CR) normalized by the root-mean-

Fig. 5. Block Diagram of OFDM System with Clipping and Filtering ( Cho et al., 2011)

(11)

(12)

(13)

represent the clipped edition of the *xp*[*m*], which can be written as follows:-

square (RMS) value σ of OFDM signal can be written as follows:-

decrease the out-of-band radiation.

or

The SLM is one of the most popular PAPR reduction techniques in the literature (Lim et al., 2005). This technique is based on the phase rotations. In SLM based OFDM (SLM-OFDM) systems, a set of *V* different data blocks are created at the transmitter representing the identical information and a data block with minimum PAPR is selected for the transmission. Fig.6 shows the general block diagram of the SLM-OFDM system. Every data block is multiplied with the *V* dissimilar phase sequences, each of length *N*, *B*(*<sup>v</sup>*) = [*bv,0, bv,1, …, bv,N-1*]*T*, *v= 1, 2…V,* which results in the changed data blocks. Now suppose the altered data block for the *v*th phase sequence is given by *X*(*<sup>v</sup>*) = [*X0bv,0, X1bv,1*,…, *XN-1bv,N-1*]*T, v=1, 2… V.* Each *<sup>v</sup> Xn* can be defined as follows:-

$$X\_n^\upsilon = X\_n b\_{\upsilon, n} \quad , \quad (1 \le \upsilon \le \mathcal{V}) \tag{14}$$

After applying SLM to *X*, the OFDM signal becomes as follows:-

$$\mathbf{x}\_{n}^{\{\upsilon\}} = \frac{1}{\sqrt{N}} \sum\_{k=0}^{N-1} X\_{k}^{\upsilon}. \ e^{j2\pi \frac{n}{N}k}, \ \mathbf{n=0, 1, 2... N-1} \tag{15}$$

where*, v = 1, 2… V.* Amongst all the tailored data blocks: *x*(*<sup>v</sup>*), *v = 1, 2… V,* the data block with minimum PAPR is selected for the transmission. Side information about the selected phase sequence must be communicated to the receiver which performs the reverse operation in order to recover the actual data block.

Fig. 6. Block diagram of OFDM system with Selective Mapping ( Han & Lee, 2005)

On PAPR Reduction Techniques in Mobile WiMAX 199

Therefore, we should execute a complete search for (*M*-1) phase factors. So, to find the optimum set of phase factors the *WM-1* sets of phase factors are searched. If we increase the number of sub-blocks *M*, the search complexity is increases exponentially. PTS needs *M*  IFFT operations for every data block, and the number of needed side information bits is [log2WM-1]. The amount of PAPR reduction is based on the number of sub-blocks *M* and the number of permitted phase factors *W*. Subblock partitioning is another factor that may have an effect on the PAPR gain PTS, which is the way of partition of the subcarriers into several disjoint sub-blocks. There are three kinds of sub-block partitioning techniques: interleaved,

Among them, pseudo-random partitioning has been found to be the best choice for PTS. The PTS technique can work with a random number of subcarriers and any modulation scheme. As mentioned above, the ordinary PTS technique has exponentially increasing search complexity. To lower the search complexity, a range of techniques have been proposed in the literature. Once the PAPR falls below a set threshold, the iterations for updating the set of phase factors must be stopped. Number of techniques has been presented in the literature to reduce the number of iterations. These techniques achieve considerable reduction in

*Example*: The PTS PAPR reduction technique for an OFDM system can be explained with a simple example ( Han & Lee, 2005). Here, we take eight subcarriers that are divided into four sub-blocks. The phase factors are selected in *P* ={±1}. Fig.8 illustrates the adjacent subblock partitioning for a data block *X* of size 8. The original data block *X* has a PAPR of 6.5

Fig. 8. An example of adjacent subblock partitioning in PTS ( Han & Lee, 2005)

Amongst them *<sup>T</sup> <sup>T</sup> bbbb* ]1,1,1,1[] <sup>ˆ</sup> , <sup>ˆ</sup> , <sup>ˆ</sup> , <sup>ˆ</sup> [ <sup>4321</sup> gets the lower PAPR. The tailored data

PAPR gain. In this case, the number of necessary IFFT operations is 4 and the amount of side information is 3 bits. The side information should be transmitted to the receiver for the recovery of original data block. There are many ways to transmit side information; one of

whose PAPR is 2.2 dB, resulting in a 4.3 dB

pseudo-random and adjacent partitioning.

search complexity with minor PAPR performance degradation.

dB. There are 8 ways to mix the sub-blocks with fixed *b1* =1.

block will be [1,-1,-1,1,-1,1,1,1] *<sup>M</sup> m m*

*m 1 X b .X ˆ* 

#### **3.3 Partial Transmit Sequence (PTS)**

PTS is another very popular PAPR reduction technique (Han & Lee, 2004; Müller & Huber, 1997; Cimini & Sollenberger, 2000; Tellambura, 2001). In this technique, the input data block of *N* symbols is partitioned into disjoint sub-blocks. In each sub-block, the subcarriers are weighted by a phase factor. The phase factors are chosen in such a way so that the PAPR of the combined signal is reduced. Fig.7 shows the general block diagram of the PTS PAPR reduction technique. In the PTS technique input data block *X* is partitioned into *M* disjoint

$$\text{1.sub-blocks: } X^{m} = [X^{m,0}, X^{m,1}, \dots, X^{m,N-1}]^{\mathsf{T}}, \ m = \text{1.2}, \mathfrak{Z}, \dots, M, \text{ such that } \sum\_{m=1}^{M} X^{m} = X \text{ and the sub-sub-shift } \mathfrak{Z} \text{ is a basis of } \mathfrak{Z} \text{ with } \mathfrak{Z} \text{ is a basis of } \mathfrak{Z}$$

blocks are combined to reduce the PAPR in the time-domain. The *L*-times oversampled time-domain signal of *Xm*, *m*=1,2,3,…,*M* is denoted by: *xm* =[*xm,0,xm,1,…,xm,N-1*]*T*, *m*=1,2,3,…,*M* is obtained by obtained by taking an IFFT of length *NL* on *Xm* concatenated with (*L*-1)*N* zeros. These are called PTS. Complex phase factors *bm* =exp(*jΦm*), *m*=1,2,3,… *M*, are launched to combine the PTSs. The set of phase factors is designated as a vector: *ˆ ˆˆ ˆ 1 2 MT b [b ,b ,...,b ]* . The time domain-signal after combining can be written as follows:-

$$\boldsymbol{\dot{\alpha}} = \sum\_{m=1}^{M} \boldsymbol{\hat{b}}^{m}. \boldsymbol{\hat{\boldsymbol{\omega}}}^{m} \tag{16}$$

The key idea is to find out the set of phase factors that reduces the PAPR. Generally, to reduce the search complexity, the selection of the phase factors is bounded by a set with a finite number of elements. The set of acceptable phase factors can be written as: *<sup>l</sup> j2 P { e : l 0,1,2,...,W 1} <sup>W</sup>* , where *W* is the number of permitted phase factors.

Additionally, we can set *b1* =1 without any loss of the performance.

Fig. 7. Block diagram of OFDM system with Partial Transmit Sequence (Müller & Huber, 1997)

PTS is another very popular PAPR reduction technique (Han & Lee, 2004; Müller & Huber, 1997; Cimini & Sollenberger, 2000; Tellambura, 2001). In this technique, the input data block of *N* symbols is partitioned into disjoint sub-blocks. In each sub-block, the subcarriers are weighted by a phase factor. The phase factors are chosen in such a way so that the PAPR of the combined signal is reduced. Fig.7 shows the general block diagram of the PTS PAPR reduction technique. In the PTS technique input data block *X* is partitioned into *M* disjoint

blocks are combined to reduce the PAPR in the time-domain. The *L*-times oversampled time-domain signal of *Xm*, *m*=1,2,3,…,*M* is denoted by: *xm* =[*xm,0,xm,1,…,xm,N-1*]*T*, *m*=1,2,3,…,*M* is obtained by obtained by taking an IFFT of length *NL* on *Xm* concatenated with (*L*-1)*N* zeros. These are called PTS. Complex phase factors *bm* =exp(*jΦm*), *m*=1,2,3,… *M*, are launched to combine the PTSs. The set of phase factors is designated as a vector: *ˆ ˆˆ ˆ 1 2 MT b [b ,b ,...,b ]* . The time domain-signal after combining can be written as

The key idea is to find out the set of phase factors that reduces the PAPR. Generally, to reduce the search complexity, the selection of the phase factors is bounded by a set with a finite number of elements. The set of acceptable phase factors can be written as:

, where *W* is the number of permitted phase factors.

Fig. 7. Block diagram of OFDM system with Partial Transmit Sequence (Müller & Huber, 1997)

*<sup>M</sup> <sup>m</sup> m 1*

*X X*

and the sub-

(16)

sub-blocks: *Xm* =[*Xm,0,Xm,1,…,Xm,N-1*]*T*, *m* =1,2,3,…,*M*, such that

Additionally, we can set *b1* =1 without any loss of the performance.

**3.3 Partial Transmit Sequence (PTS)** 

follows:-

*<sup>l</sup> j2*

*P { e : l 0,1,2,...,W 1} <sup>W</sup>*

Therefore, we should execute a complete search for (*M*-1) phase factors. So, to find the optimum set of phase factors the *WM-1* sets of phase factors are searched. If we increase the number of sub-blocks *M*, the search complexity is increases exponentially. PTS needs *M*  IFFT operations for every data block, and the number of needed side information bits is [log2WM-1]. The amount of PAPR reduction is based on the number of sub-blocks *M* and the number of permitted phase factors *W*. Subblock partitioning is another factor that may have an effect on the PAPR gain PTS, which is the way of partition of the subcarriers into several disjoint sub-blocks. There are three kinds of sub-block partitioning techniques: interleaved, pseudo-random and adjacent partitioning.

Among them, pseudo-random partitioning has been found to be the best choice for PTS. The PTS technique can work with a random number of subcarriers and any modulation scheme. As mentioned above, the ordinary PTS technique has exponentially increasing search complexity. To lower the search complexity, a range of techniques have been proposed in the literature. Once the PAPR falls below a set threshold, the iterations for updating the set of phase factors must be stopped. Number of techniques has been presented in the literature to reduce the number of iterations. These techniques achieve considerable reduction in search complexity with minor PAPR performance degradation.

*Example*: The PTS PAPR reduction technique for an OFDM system can be explained with a simple example ( Han & Lee, 2005). Here, we take eight subcarriers that are divided into four sub-blocks. The phase factors are selected in *P* ={±1}. Fig.8 illustrates the adjacent subblock partitioning for a data block *X* of size 8. The original data block *X* has a PAPR of 6.5 dB. There are 8 ways to mix the sub-blocks with fixed *b1* =1.

Fig. 8. An example of adjacent subblock partitioning in PTS ( Han & Lee, 2005)

Amongst them *<sup>T</sup> <sup>T</sup> bbbb* ]1,1,1,1[] <sup>ˆ</sup> , <sup>ˆ</sup> , <sup>ˆ</sup> , <sup>ˆ</sup> [ <sup>4321</sup> gets the lower PAPR. The tailored data block will be [1,-1,-1,1,-1,1,1,1] *<sup>M</sup> m m m 1 X b .X ˆ* whose PAPR is 2.2 dB, resulting in a 4.3 dB PAPR gain. In this case, the number of necessary IFFT operations is 4 and the amount of side information is 3 bits. The side information should be transmitted to the receiver for the recovery of original data block. There are many ways to transmit side information; one of

On PAPR Reduction Techniques in Mobile WiMAX 201

Fig. 9. Block diagram of OFDM system with Precording Techniques ( Baig & Jeoti, 2010)

matrix of size *N=L×L* and *Ym* can be written as follows:-

written as:-

Fig.9 shows the precoding based OFDM system. In these system, the kernel of the WHT/DHT/DCT acts as a precoding matrix *P* of dimension *N= L×L* and it is applied to constellations symbols before the IFFT to reduce the correlation among the input sequence. In the precoding based systems baseband modulated data is passed through S/P converter which generates a complex vector of size *L* that can be written as *X*=[*X0, X1, …, XL-1*]T .Then precoding is applied to this complex vector which transforms this complex vector into new vector of length *L* that can be written as Y=PX=[*Y0, Y1, …, YL-1*]T, where *P* is a precoder

 (24) *Pm,l* means *mth* row and *lth* column of precoder matrix. Equation (24) represents the precoded constellations symbols. The complex baseband OFDM signal with *N* subcarriers can be

Table 1 summarizes the PAPR reduction techniques presented in the literature review. The clipping techniques have low implementation complexity but on the other hand, the clipping operation may introduce both in-band distortion and out-of-band radiation into the multicarrier signals, which degrades the OFDM system performance including BER and spectral efficiency. SLM and PTS both have high computational complexity. However, the

and DCT can be defined as:-

(22)

(23)

(25)

the ways is to transmit side information bits with a separate channel other than the data channel. Another ways is to include the side information within the data block but it results in data rate loss.

#### **3.4 Precoding based techniques**

Precoding based techniques are simple linear techniques. These techniques can reduce the PAPR up to the PAPR of single carrier systems (Slimane, 2007). Walsh-Hadamard transform (WHT) precoding based techniques, discrete cosine transform (DCT) precoding based techniques, discrete hartley transform (DHT) precoding based techniques are common examples of precoding based PAPR reduction techniques (Slimane, 2007; Min & Jeoti, 2007; Baig & Jeoti, 2010a, 2010b, 2010c).

#### **3.4.1 Walsh-Hadamard Transform (WHT)**

WHT is an orthogonal linear transform and can be implemented by a butterfly structure as in FFT. This means that applying WHT does not require the extensive increase of system complexity. The kernel of WHT can be written as follows:-

$$H\_1 = \lfloor 1 \rfloor \tag{17}$$

$$H\_2 = \frac{1}{\sqrt{2}} \begin{bmatrix} \mathbf{1} & & \mathbf{1} \\ \mathbf{1} & -\mathbf{1} \end{bmatrix} \tag{18}$$

$$H\_{2N} = \frac{1}{\sqrt{2N}} \begin{bmatrix} H\_N & H\_N \\ H\_N & H\_N^{-1} \end{bmatrix} \tag{19}$$

where *<sup>1</sup> HN* denotes the binary complement of *HN*.

#### **3.4.2 Discrete Hartley Transform (DHT)**

DHT is a linear transform. In DHT *N* real numbers *x0, x1, …, xN-1* are transformed in to *N* real numbers *H0, H1, …, HN-1*. The *N*-point DHT can be defined as follows:-

$$\begin{split} H\_k &= \Sigma\_{n=0}^{N=1} \chi\_n \left[ \cos \left( \frac{2\pi nk}{N} \right) + \sin \left( \frac{2\pi nk}{N} \right) \right] \\ &= \Sigma\_{n=0}^{N=1} \chi(n) . \cos \left( \frac{2\pi nk}{N} \right) \end{split} \tag{20}$$

$$p\_{m,n} = \cos\left(\frac{2\pi mn}{N}\right) \tag{21}$$

Where *cas cos sin* and *k*=1, 2, 3… *N*-1.The DHT is also invertible transform which allows us to recover the *xn* from *Hk* and inverse can be obtained by simply multiplying DHT of *Hk* by *<sup>1</sup> N* .

#### **3.4.3 Discrete Cosine Transform (DCT)**

DCT matrix *P* of size *N*-by-*N* can be created by using equation (22)

$$\mathcal{D}\_{ij} = \begin{cases} \frac{1}{\sqrt{N}} & \mathfrak{i} = \mathbf{0}, \qquad \mathbf{0} \le \mathfrak{j} \le N - \mathbf{1} \\\\ \frac{2}{N} \cos \frac{\pi (2j + \mathbf{1}) \mathfrak{i}}{2N} & \mathbf{1} \le \mathfrak{i} \le N - \mathbf{1} \\\\ \mathbf{0} \le j \le N - \mathbf{1} \end{cases} \tag{22}$$

and DCT can be defined as:-

200 Advanced Transmission Techniques in WiMAX

the ways is to transmit side information bits with a separate channel other than the data channel. Another ways is to include the side information within the data block but it results

Precoding based techniques are simple linear techniques. These techniques can reduce the PAPR up to the PAPR of single carrier systems (Slimane, 2007). Walsh-Hadamard transform (WHT) precoding based techniques, discrete cosine transform (DCT) precoding based techniques, discrete hartley transform (DHT) precoding based techniques are common examples of precoding based PAPR reduction techniques (Slimane, 2007; Min & Jeoti, 2007;

WHT is an orthogonal linear transform and can be implemented by a butterfly structure as in FFT. This means that applying WHT does not require the extensive increase of system

DHT is a linear transform. In DHT *N* real numbers *x0, x1, …, xN-1* are transformed in to *N* real

Where *cas cos sin* and *k*=1, 2, 3… *N*-1.The DHT is also invertible transform which allows us to recover the *xn* from *Hk* and inverse can be obtained by simply multiplying DHT

(17)

(18)

(19)

(20)

(21)

in data rate loss.

where *<sup>1</sup> HN*

of *Hk* by *<sup>1</sup>*

*N* .

**3.4 Precoding based techniques** 

Baig & Jeoti, 2010a, 2010b, 2010c).

**3.4.1 Walsh-Hadamard Transform (WHT)** 

**3.4.2 Discrete Hartley Transform (DHT)** 

**3.4.3 Discrete Cosine Transform (DCT)** 

complexity. The kernel of WHT can be written as follows:-

denotes the binary complement of *HN*.

numbers *H0, H1, …, HN-1*. The *N*-point DHT can be defined as follows:-

DCT matrix *P* of size *N*-by-*N* can be created by using equation (22)

(23)

Fig. 9. Block diagram of OFDM system with Precording Techniques ( Baig & Jeoti, 2010)

Fig.9 shows the precoding based OFDM system. In these system, the kernel of the WHT/DHT/DCT acts as a precoding matrix *P* of dimension *N= L×L* and it is applied to constellations symbols before the IFFT to reduce the correlation among the input sequence. In the precoding based systems baseband modulated data is passed through S/P converter which generates a complex vector of size *L* that can be written as *X*=[*X0, X1, …, XL-1*]T .Then precoding is applied to this complex vector which transforms this complex vector into new vector of length *L* that can be written as Y=PX=[*Y0, Y1, …, YL-1*]T, where *P* is a precoder matrix of size *N=L×L* and *Ym* can be written as follows:-

$$Y\_m = \sum\_{l=0}^{L-1} p\_{m,l} \cdot X\_l \qquad m = 0, 1, \dots \\ L - 1 \tag{24}$$

*Pm,l* means *mth* row and *lth* column of precoder matrix. Equation (24) represents the precoded constellations symbols. The complex baseband OFDM signal with *N* subcarriers can be written as:-

$$\alpha\_n = \frac{1}{\sqrt{N}} \begin{array}{c} \sum\_{m=0}^{N-1} Y\_m \ . \ e^{j2\pi \frac{n}{N}m} \end{array} \qquad \qquad n = 0, 1, 2\dots N-1 \tag{25}$$

Table 1 summarizes the PAPR reduction techniques presented in the literature review. The clipping techniques have low implementation complexity but on the other hand, the clipping operation may introduce both in-band distortion and out-of-band radiation into the multicarrier signals, which degrades the OFDM system performance including BER and spectral efficiency. SLM and PTS both have high computational complexity. However, the

On PAPR Reduction Techniques in Mobile WiMAX 203

Let us consider the periodic correlation property of Zadoff-Chu sequences with the same

For the sake of simplicity we put *q*=0 and *r*=1 in equation (26), then using equation (26) in

Where *k* =0,1,2,…,*N*-1 , *q* is any integer, *r* is any integer relatively prime to *N*.

prime length. The periodic cross-correlation function can be defined as follows:-

While *N* is even 2*m*-2*n*-*N* is even. So, the equation (29) can be stated as: exp(

From equation (30), it is obvious that *ρ*(*m*) =0, when *m* =0. Since, *m* and *N* are relatively

unity but not equal to 1 for the range of *m* {*m*:1,2,3,…,*N*} shown in equation (30). Hence, we

where *r* is *Nth* root of unity, substituting equation (31) into equation (30), we get *(m) 0,{m : 1,2,3,...,N }* , At the end, it is concluded that the ideal periodic autocorrelation

is a primitive *Nth* root of unity. Therefore,

Hence, we can combine two summations of equation (28) as follows:-

**4.2 Autocorrelation property** 

the above expression we get:-

prime to each other,

*j2 m [ ] <sup>N</sup> <sup>e</sup>* 

can employ the theorem as follows:

(26)

(27)

(28)

(29)

(30)

(31)

*<sup>2</sup> j (2mn m ) N* ).

is a *Nth* root of

*j2 m [ ] <sup>N</sup> <sup>e</sup>* 


precoding based techniques show great promise as they are simple linear techniques to implement without any complex optimizations.

Table 1. Comparison of the PAPR Reduction Techniques

The main characteristics of precoding based techniques are, no bandwidth expansion, no power increase, no data rate loss, no BER degradation and distortionless.

#### **4. Proposed PAPR reduction techniques**

The random interleaved subcarrier mapping is favourable for the mobile WiMAX because it increases the capacity in the frequency selective fading channels and offers maximum frequency diversity. So, in this section, we present two precoding based random-interleaved OFDMA uplink systems for the PAPR reduction in the mobile WiMAX systems: Zadoff-Chu matrix transform (ZCMT) precoding based random-interleaved OFDMA uplink system and SLM based ZCMT precoded random-interleaved OFDMA uplink system. The PAPR of the proposed system is analyzed with root-raised-cosine (RRC) pulse shaping.

#### **4.1 Zadoff-Chu sequences**

Zadoff-Chu sequences are the class of polyphase sequences having optimum correlation properties. These sequences have an ideal periodic autocorrelation, constant magnitude and circular auto-orthogonality. The constant envelope feature of the Zadoff-Chu sequences can greatly alleviate the annoying peak-to-average power (PAPR) problem occurred in orthogonal frequency division multiplexing (OFDM) systems. According to (Chu, 1972; Popovic´, 1997), Zadoff-Chu sequences of length N can be defined as follows:-

$$z\_n = \begin{cases} \begin{array}{c} e^{\frac{j2\pi r}{N} \left(\frac{k^2}{\pi} + qk\right)} & \text{for } N \to \infty \\\\ \begin{array}{c} j\frac{2\pi r}{N} \left(\frac{k(\&\*\,\*)}{\pi} + qk\right) \end{array} . \end{cases} . \tag{26}$$

Where *k* =0,1,2,…,*N*-1 , *q* is any integer, *r* is any integer relatively prime to *N*.

#### **4.2 Autocorrelation property**

202 Advanced Transmission Techniques in WiMAX

precoding based techniques show great promise as they are simple linear techniques to

**Distortion BER** 

Filtering LOW YES YES NO NO NO

Mapping HIGH NO NO YES NO NO

The main characteristics of precoding based techniques are, no bandwidth expansion, no

The random interleaved subcarrier mapping is favourable for the mobile WiMAX because it increases the capacity in the frequency selective fading channels and offers maximum frequency diversity. So, in this section, we present two precoding based random-interleaved OFDMA uplink systems for the PAPR reduction in the mobile WiMAX systems: Zadoff-Chu matrix transform (ZCMT) precoding based random-interleaved OFDMA uplink system and SLM based ZCMT precoded random-interleaved OFDMA uplink system. The PAPR of the

Zadoff-Chu sequences are the class of polyphase sequences having optimum correlation properties. These sequences have an ideal periodic autocorrelation, constant magnitude and circular auto-orthogonality. The constant envelope feature of the Zadoff-Chu sequences can greatly alleviate the annoying peak-to-average power (PAPR) problem occurred in orthogonal frequency division multiplexing (OFDM) systems. According to (Chu, 1972;

**Degradation**

HIGH NO NO YES NO YES

LOW NO NO NO NO NO

**Bandwidth Expansion**

**Power Increase**  **Data Rate Loss** 

implement without any complex optimizations.

**Implementation Complexity** 

Table 1. Comparison of the PAPR Reduction Techniques

**4. Proposed PAPR reduction techniques** 

**4.1 Zadoff-Chu sequences** 

power increase, no data rate loss, no BER degradation and distortionless.

proposed system is analyzed with root-raised-cosine (RRC) pulse shaping.

Popovic´, 1997), Zadoff-Chu sequences of length N can be defined as follows:-

**PAPR Reduction Technique** 

Clipping &

Selective

Partial Transmit Sequence

Precoding Based Techniques Let us consider the periodic correlation property of Zadoff-Chu sequences with the same prime length. The periodic cross-correlation function can be defined as follows:-

$$\begin{array}{l} \rho(m) = \sum\_{n=0}^{N-1} z\_n z\_{(n-m)modN}^\* \\ = \sum\_{n=0}^{m-1} z\_n z\_{(n-m+N)}^\* + \sum\_{n=m}^{N-1} z\_n z\_{(n-m)}^\* \end{array} \tag{27}$$

For the sake of simplicity we put *q*=0 and *r*=1 in equation (26), then using equation (26) in the above expression we get:-

$$\begin{array}{l} = \sum\_{n=0}^{m-1} e^{\frac{\lfloor j\pi n^{2} \rfloor}{N}} \cdot e^{\frac{\lfloor -j\pi (n-m+N)^{2} \rfloor}{N}} + \sum\_{n=m}^{N-1} e^{\frac{\lfloor j\pi n^{2} \rfloor}{N}} \cdot e^{\frac{\lfloor -j\pi (n-m)^{2} \rfloor}{N}}\\ = \sum\_{n=0}^{m-1} e^{\frac{\lfloor j\pi (2mn-n2N-m^{2}+2mnN+N^{2}) \rfloor}{N}} + \sum\_{n=m}^{N-1} e^{\frac{\lfloor j\pi (n^{2}-n^{2}-m^{2}+2mnN) \rfloor}{N}}\\ = \sum\_{n=0}^{m-1} e^{\frac{\lfloor j\pi (2mn-m+N)(m-N) \rfloor}{N}} + \sum\_{n=m}^{N-1} e^{\frac{\lfloor j\pi (2mn-n^{2}) \rfloor}{N}} \end{array} \tag{28}$$

$$\begin{aligned} \text{Note that,} \qquad \begin{aligned} \text{e}^{\frac{\lfloor \pi(2m-m+N)/(m-N) \rfloor}{N}} &= \text{e}^{\lfloor \frac{\pi(2mn-m+2N-m^2+\pi mN+N^2)}{N} \rfloor} \\ &= \text{e}^{\lfloor \frac{\pi(2mn-m^2)}{N} \rfloor} . \text{e}^{\lfloor \frac{\pi(2mn-\pi m^2)}{N} \rfloor} \\ &= \text{e}^{\lfloor \frac{\pi(2mn-m^2)}{N} \rfloor} . \text{e}^{\lfloor \pi(2m-2n-N) \rfloor} \end{aligned} \tag{29}$$

While *N* is even 2*m*-2*n*-*N* is even. So, the equation (29) can be stated as: exp( *<sup>2</sup> j (2mn m ) N* ). Hence, we can combine two summations of equation (28) as follows:-

$$\begin{array}{l} \rho(m) = \Sigma\_{n=0}^{N-1} e^{\frac{\left[j\pi(zmn - m^2)\right]}{N}} \\ = e^{\left[\frac{-j\pi m^2}{N}\right]} \cdot \Sigma\_{n=0}^{N-1} e^{\left[\frac{j\pi \pi mn}{N}\right]} \end{array} \tag{30}$$

From equation (30), it is obvious that *ρ*(*m*) =0, when *m* =0. Since, *m* and *N* are relatively prime to each other, *j2 m [ ] <sup>N</sup> <sup>e</sup>* is a primitive *Nth* root of unity. Therefore, *j2 m [ ] <sup>N</sup> <sup>e</sup>* is a *Nth* root of unity but not equal to 1 for the range of *m* {*m*:1,2,3,…,*N*} shown in equation (30). Hence, we can employ the theorem as follows:

$$
\Sigma\_{n=0}^{N-1} r^n = \begin{cases} N & r = 1 \\ 0 & r \neq 1 \end{cases} \tag{31}
$$

where *r* is *Nth* root of unity, substituting equation (31) into equation (30), we get *(m) 0,{m : 1,2,3,...,N }* , At the end, it is concluded that the ideal periodic autocorrelation

On PAPR Reduction Techniques in Mobile WiMAX 205

Fig. 10. ZCMT Precording based Random Interleaved OFDMA Uplink System

where *Z* is a precoder matrix of size *N*=*L*×*L* and *Ym* can be written as follows:-

vector into new vector of same length *L* that can be written as follows:-

reshaping *k* =*mL*+*l* in equation (26) we get:-

Since, *N*=*L*2, then equation (37) can be reduced to

Then ZCMT precoding is applied to this complex vector which transforms this complex

*zl,m* means *lth* row of precoder matrix. Expanding equation (36), using row wise sequence

(38)

(34)

(35)

(36)

(37)

property of Zadoff-Chu sequences makes it suitable candidate for PAPR reduction in OFDM systems.

#### **4.3 Constant envelope property after IDFT**

The Zadoff-Chu sequences have constant amplitude, and its IDFT has also constant amplitude. Additionally, zadoff-chu sequence is also Zadoff-Chu sequence after FFT or IFFT.

#### **4.4 Orthogonality property**

The DFT of the Zadoff-Chu sequences equal to the conjugate of the Zadoff-chu sequences as follows:-

$$\text{DFT}(\mathbf{z}\_n) = (\mathbf{z}\_n)^\* \tag{32}$$

Therefore, the orthogonality in time domain as well as in frequency domain is preserved.

#### **4.5 Zadoff-Chu Matrix Transform (ZCMT)**

Zadoff-Chu matrix transform (ZCMT) is used to lower the correlation relationship of the IFFT input sequence. The ZCMT precoding matrix must accomplish the following criteria:-


The first condition guarantees that each output symbol has the same quantity of information of every input data. The second requirement preserves the power at the precoder output. Finally, the third requirement ensures the recovery of the original data at the receiver. The kernel of the ZCMT is defined in equation (26). For *N* =*L*×*L* and *j 1* , the ZCMT kernel *Z*, of size *N* = *L*×*L*=*L2* is obtained by reshaping the Zadoff-Chu sequence row-wise by *k* =*mL*+*l*, as hereunder:-

$$Z = \frac{1}{\sqrt{N}} \begin{bmatrix} Z\_{00} & Z\_{01} & \dots & Z\_{0(L-1)} \\ Z\_{10} & Z\_{11} & \dots & Z\_{1(L-1)} \\ \vdots & \vdots & \ddots & \vdots \\ Z\_{(L-1)0} & Z\_{(L-1)1} & \dots & Z\_{(L-1)(L-1)} \end{bmatrix} \tag{33}$$

In other words, the *L2* point long Zadoff-Chu sequence fills the kernel of the matrix transform row-wise.

#### **4.6 Proposed ZCMT precoding based random interleaved OFDMA system**

Fig.10 shows a ZCMT precoding based random-interleaved OFDMA uplink system. In this system a precoding matrix *Z* of dimension *N*=*L*×*L* is applied to constellations symbols before the subcarrier mapping and IFFT to reduce the PAPR. In the ZCMT precoding based random-interleaved OFDMA systems, baseband modulated data is passed through S/P convertor which generates a complex vector of size *M* that can be written as follows:-

property of Zadoff-Chu sequences makes it suitable candidate for PAPR reduction in OFDM

The Zadoff-Chu sequences have constant amplitude, and its IDFT has also constant amplitude. Additionally, zadoff-chu sequence is also Zadoff-Chu sequence after FFT or

The DFT of the Zadoff-Chu sequences equal to the conjugate of the Zadoff-chu sequences as

Therefore, the orthogonality in time domain as well as in frequency domain is preserved.

Zadoff-Chu matrix transform (ZCMT) is used to lower the correlation relationship of the IFFT input sequence. The ZCMT precoding matrix must accomplish the following criteria:-

The first condition guarantees that each output symbol has the same quantity of information of every input data. The second requirement preserves the power at the precoder output. Finally, the third requirement ensures the recovery of the original data at the receiver. The kernel of the ZCMT is defined in equation (26). For *N* =*L*×*L* and *j 1* , the ZCMT kernel *Z*, of size *N* = *L*×*L*=*L2* is obtained by reshaping the Zadoff-Chu sequence row-wise by *k*

In other words, the *L2* point long Zadoff-Chu sequence fills the kernel of the matrix

Fig.10 shows a ZCMT precoding based random-interleaved OFDMA uplink system. In this system a precoding matrix *Z* of dimension *N*=*L*×*L* is applied to constellations symbols before the subcarrier mapping and IFFT to reduce the PAPR. In the ZCMT precoding based random-interleaved OFDMA systems, baseband modulated data is passed through S/P

convertor which generates a complex vector of size *M* that can be written as follows:-

**4.6 Proposed ZCMT precoding based random interleaved OFDMA system** 

1. All the elements of the precoding matrix must have the identical magnitude.

*N* .

(32)

(33)

systems.

IFFT.

follows:-

**4.3 Constant envelope property after IDFT** 

**4.5 Zadoff-Chu Matrix Transform (ZCMT)** 

2. The magnitude must be equal to: *<sup>1</sup>*

=*mL*+*l*, as hereunder:-

transform row-wise.

3. The ZCMT precoding matrix must be non-singular.

**4.4 Orthogonality property** 

Fig. 10. ZCMT Precording based Random Interleaved OFDMA Uplink System

Then ZCMT precoding is applied to this complex vector which transforms this complex vector into new vector of same length *L* that can be written as follows:-

$$Y = ZX = [Y\_0, Y\_1, Y\_2 \dots Y\_{M-1}]^T \tag{35}$$

where *Z* is a precoder matrix of size *N*=*L*×*L* and *Ym* can be written as follows:-

$$Y\_l = \Sigma\_{m=0}^{L-1} z\_{l,m} \cdot X\_m \qquad l = 0, 1, \ldots \\ L - 1 \tag{36}$$

*zl,m* means *lth* row of precoder matrix. Expanding equation (36), using row wise sequence reshaping *k* =*mL*+*l* in equation (26) we get:-

$$\begin{split} Y\_{1} = \frac{1}{\sqrt{N}} \sum\_{m=0}^{L-1} \text{(e} \begin{pmatrix} \frac{j\pi rl}{N} \Big( \frac{(mL+l)^{2}}{2} + q(mL+l) \Big) \\ \end{pmatrix} . X\_{m} &= \frac{1}{\sqrt{N}} \sum\_{m=0}^{L-1} \text{(e} \begin{vmatrix} \frac{j\pi rl}{N} \Big( \frac{j\pi rl}{2} \Big( \frac{j\pi rl}{2} \Big) & \frac{j\pi rglL}{2} \Big) \\ \end{vmatrix} . X\_{m} \\ &= \frac{1}{\sqrt{N}} \sum\_{m=0}^{L-1} \text{(e} \begin{vmatrix} \frac{j\pi rm^{2}L^{2}}{N} & \frac{j\pi rl^{2}}{N} \int \frac{j\pi rrmlL}{N} \cdot \frac{j\pi rrqmlL}{2} \cdot \frac{j\pi rrqml}{N} & \frac{j\pi rgl}{N} \\ \end{vmatrix} . \end{split} \tag{37}$$

Since, *N*=*L*2, then equation (37) can be reduced to

$$\hat{\mathbf{r}} = \frac{1}{L} .e^{\frac{j\pi r l^2}{L^2}} .e^{\frac{j\pi \pi r lq}{L^2}} \sum\_{m=0}^{j\pi \pi r lq} (e^{\frac{j\pi \pi r l m l}{L}} . \mathcal{X}\_m) \tag{38}$$

(34)

On PAPR Reduction Techniques in Mobile WiMAX 207

*Re(c) c , c c n n* .

, *m* =0,1,2,…,*N*-1

is the aperiodic autocorrelation function. It is concluded from equation (46) that if the aperiodic autocorrelation mold of the IFFT input sequence *xz* is small (*ρ*(*m*) for ≥1 ) then, the peak-power factor of the signal obtained by passing through the OFDM multi-carrier system also can be small (Tellambura, 1997).The peak value of the autocorrelation is the averagepower of the input sequence. After that, if the number of subcarriers is not altered, this peak-value completely depends on the input sequence. It means that if the sidelobe of an autocorrelation function of an input sequence has greater value than other input sequences, the former has high correlation property. The IFFT operation can be expressed as multiplying sinusoidal functions to the input sequence, summing and sampling the results. Hence, the high correlation property of the IFFT input causes the sinusoidal functions to be arranged with in-phase form. After summing these in-phase functions, the output might

To verify the contribution of ZCMT, we consider OFDMA system for QPSK modulation. Fig.11 shows that the aperiodic autocorrelation function of randomly generated QPSK sequence with the length 64 is given, which are normalized by the length. Thus the maximum value is 1 which is the average power of the sequence. It is obvious from the Fig.11 that two autocorrelation functions have different sidelobe value. If the sidelobes of autocorrelation have higher values, then the input sequence is highly correlated and its PAPR is high. The high correlation in the input to IFFT causes the subcarriers to align inphase. After summing these in-phase functions, the output might have high amplitude resulting in higher PAPR. The sidelobe value of the proposed ZCMT is much smaller than the conventional OFDMA systems. Therefore, it is concluded that if we apply ZCMT precoder to the IFFT input sequence, it lower the correlation relationship of the OFDMA

*N1m \* zi z(i m) i 0 (m) x x*

For any complex c,

That's why,

have large peaks.

**4.6.1 The effect of Zadoff-Chu Matrix Transform** 

input sequence, thus PAPR can be reduced.

where,

(44)

(45)

(46)

where, *2 2 2 j rm j2 rml L L X (e .e ).X , m m m*=0,1,2,…,*L*-1, *l* =0,1,2,…,*L*-1. Equation (38) represents the ZCMT precoded constellations symbols. After precoding operation, the subcarrier mapping is performed on these ZCMT precoded constellations symbols in random-interleaved mode. After the subcarrier mapping in random interleaved mode, we get frequency domain samples 0,1,2,..., 1 *<sup>l</sup> ˆ {Y : l N }.* . Mathematically, the subcarrier mapping in random interleaved mode can be done as follows:-

$$\hat{\mathcal{Y}}\_{l} = \begin{cases} Y\_{l} & , l = \hat{Q} \cdot k \\ \mathbf{0} & , \end{cases} \quad \mathbf{0} \le k \le M - 1 \\ \tag{39}$$

Where *ˆ 0 l N 1,N Q.M,0 Q Q* , *N*: System subcarriers, *M*: User subcarriers (for one user), *Q* : Subchannels/Users (*Q*=*N*/*L*) . The *kth* subcarrier of each group is assigned to the *kth* user with index set: *q ,1 q ,2 q,L 1 {( ),(Q ),...,(L 1)Q )}* , where *<sup>q</sup> ,1 <sup>q</sup> ,2 <sup>q</sup> ,L 1 {( ),( ),...,( )}* are independent and identically distributed random variables with uniform distribution on (*q*=0,1,2,…,*Q*-1). Suppose the *kth* user is assigned to subchannel *k* then the complex baseband ZCMT precoded random interleaved OFDMA signal for *kth* user can be written as:-

$$\boldsymbol{\alpha}\_{n}^{(k)} = \sum\_{l=0}^{M-1} \hat{Y}\_{l}^{(k)}.\ \mathbf{e}^{j2\pi \frac{(lQ\*\mathbf{y}\_{q,k})}{N}n},\ n = \ \mathbf{0}, 1...N-1\tag{40}$$

where users index *q* =0,1,2,…,*Q*-1 and *(k) <sup>l</sup> Yˆ* is modulated signal on subcarrier *l* for *k*th user. The complex passband signal of ZCMT precoded random-interleaved OFDMA after RRC pulse shaping can be written as follows:-

$$\mathbf{x}(t) = \mathbf{e}^{j\,\omega\_{\varepsilon}t} \sum\_{n=0}^{N-1} \boldsymbol{\chi}\_{n}^{\{k\}} \cdot \mathbf{r}(t - n\ddot{T}) \tag{41}$$

where, *c* is carrier frequency, *r(t)* is baseband pulse, *<sup>M</sup> T ( ).T <sup>N</sup>* . is compressed symbol duration after IFFT and *T* is symbol duration is seconds. The PAPR of the ZCMT precoded random-interleaved OFDMA signal in equation (41) with pulse shaping can be written as follows:-

$$PAPR(dB) = 10\log\_{10}\frac{\max\left(|\kappa(t)|^2\right)}{E\{\max\left(|\kappa(t)|^2\right)\}}\tag{42}$$

*E*{.}, denote expected value. If the amplitude of all subcarriers are normalized, *<sup>2</sup> E{max( x(t) )} N* , the equation (42) reduced to:-

$$PAPR(dB) = 10\log\_{10}\frac{\max\left(|x(t)|^2\right)}{N} \tag{43}$$

It should be pointed out that the orthogonality of the symbols after introducing precoding is maintained, as the precoding matrix is cyclic auto-orthogonal (Tasi et al., 2006). The instantaneous power of *x(t)* can be defined as follows:-

$$p(t) = |\boldsymbol{\chi}(t)|^2 = \boldsymbol{\chi}(t) \* \boldsymbol{\chi}^\*(t) \tag{44}$$

$$\begin{split} &= \frac{1}{N} \sum\_{i}^{N-1} \sum\_{k}^{N-1} \boldsymbol{\chi}\_{\overline{z}i} \boldsymbol{\chi}\_{\overline{z}k}^{\*} \mathbf{e}^{\{j2\pi(i-k)\mathbf{t}\}} \\ &= \frac{1}{N} \left[ N + 2 \text{Re} \{ \sum\_{l}^{N-2} \sum\_{k=1}^{N-1} \boldsymbol{\chi}\_{\overline{z}l} \boldsymbol{\chi}\_{\overline{z}k}^{\*} \mathbf{e}^{\{j2\pi(i-k)\mathbf{t}\}} \} \right] \\ &= 1 + \frac{2}{N} \text{Re} \{ \sum\_{m=1}^{N-1} \mathbf{e}^{\{j2\pi t\}} \sum\_{l=0}^{N-1-m} \boldsymbol{\chi}\_{\overline{z}l} \boldsymbol{\chi}\_{\overline{z}(i+m)}^{\*} \} \end{split} \tag{45}$$

For any complex c,

$$\operatorname{Re}(c) \le |c| \cdot \left| \sum c\_n \right| \le \left| \sum c\_n \right| \cdot \left| \right| $$

That's why,

$$p(t) \le 1 + \frac{2}{N} \Sigma\_{m=1}^{N-1} |\rho(m)| \tag{46}$$

where,

(39)

(42)

(43)

(40)

(41)

. is compressed symbol

*<sup>l</sup> Yˆ* is modulated signal on subcarrier *l* for *k*th user.

206 Advanced Transmission Techniques in WiMAX

ZCMT precoded constellations symbols. After precoding operation, the subcarrier mapping is performed on these ZCMT precoded constellations symbols in random-interleaved mode. After the subcarrier mapping in random interleaved mode, we get frequency domain

Where *ˆ 0 l N 1,N Q.M,0 Q Q* , *N*: System subcarriers, *M*: User subcarriers (for one user), *Q* : Subchannels/Users (*Q*=*N*/*L*) . The *kth* subcarrier of each group is assigned to the *kth* user with index set: *q ,1 q ,2 q,L 1 {( ),(Q ),...,(L 1)Q )}* , where *<sup>q</sup> ,1 <sup>q</sup> ,2 <sup>q</sup> ,L 1 {( ),( ),...,( )}* are independent and identically distributed random variables with uniform distribution on (*q*=0,1,2,…,*Q*-1). Suppose the *kth* user is assigned to subchannel *k* then the complex baseband ZCMT precoded random interleaved OFDMA signal for *kth* user

The complex passband signal of ZCMT precoded random-interleaved OFDMA after RRC

duration after IFFT and *T* is symbol duration is seconds. The PAPR of the ZCMT precoded random-interleaved OFDMA signal in equation (41) with pulse shaping can be written as

*E*{.}, denote expected value. If the amplitude of all subcarriers are normalized,

It should be pointed out that the orthogonality of the symbols after introducing precoding is maintained, as the precoding matrix is cyclic auto-orthogonal (Tasi et al., 2006). The

*m*=0,1,2,…,*L*-1, *l* =0,1,2,…,*L*-1. Equation (38) represents the

*ˆ {Y : l N }.* . Mathematically, the subcarrier mapping in random

where,

*2 2 2 j rm j2 rml L L X (e .e ).X , m m* 

interleaved mode can be done as follows:-

where users index *q* =0,1,2,…,*Q*-1 and *(k)*

pulse shaping can be written as follows:-

*<sup>2</sup> E{max( x(t) )} N* , the equation (42) reduced to:-

instantaneous power of *x(t)* can be defined as follows:-

where, *c* is carrier frequency, *r(t)* is baseband pulse, *<sup>M</sup> T ( ).T <sup>N</sup>*

samples 0,1,2,..., 1 *<sup>l</sup>*

can be written as:-

follows:-

$$\mathfrak{sp}(m) = \sum\_{i=0}^{N-1-m} \mathfrak{x}\_{zi} \mathfrak{x}\_{z(i+m)}^\* \;/\ m = 0, 1, 2, \dots, N-1$$

is the aperiodic autocorrelation function. It is concluded from equation (46) that if the aperiodic autocorrelation mold of the IFFT input sequence *xz* is small (*ρ*(*m*) for ≥1 ) then, the peak-power factor of the signal obtained by passing through the OFDM multi-carrier system also can be small (Tellambura, 1997).The peak value of the autocorrelation is the averagepower of the input sequence. After that, if the number of subcarriers is not altered, this peak-value completely depends on the input sequence. It means that if the sidelobe of an autocorrelation function of an input sequence has greater value than other input sequences, the former has high correlation property. The IFFT operation can be expressed as multiplying sinusoidal functions to the input sequence, summing and sampling the results. Hence, the high correlation property of the IFFT input causes the sinusoidal functions to be arranged with in-phase form. After summing these in-phase functions, the output might have large peaks.

#### **4.6.1 The effect of Zadoff-Chu Matrix Transform**

To verify the contribution of ZCMT, we consider OFDMA system for QPSK modulation. Fig.11 shows that the aperiodic autocorrelation function of randomly generated QPSK sequence with the length 64 is given, which are normalized by the length. Thus the maximum value is 1 which is the average power of the sequence. It is obvious from the Fig.11 that two autocorrelation functions have different sidelobe value. If the sidelobes of autocorrelation have higher values, then the input sequence is highly correlated and its PAPR is high. The high correlation in the input to IFFT causes the subcarriers to align inphase. After summing these in-phase functions, the output might have high amplitude resulting in higher PAPR. The sidelobe value of the proposed ZCMT is much smaller than the conventional OFDMA systems. Therefore, it is concluded that if we apply ZCMT precoder to the IFFT input sequence, it lower the correlation relationship of the OFDMA input sequence, thus PAPR can be reduced.

On PAPR Reduction Techniques in Mobile WiMAX 209

Fig. 12. Block diagram of SLM based ZCMT precoded random-interleaved OFDMA Uplink

The precoding based SLM technique needs *V* (dissimilar phase sequences) IFFT operations and the information bits required as side information for each data block is [log2V]. Precoding based SLM technique is applicable for any number of subcarriers and all types of modulation techniques. The PAPR reduction for precoding based SLM technique depends on the number of phase sequences *V* and the output data with lowest PAPR is selected by the transmitter for transmissions. The complex passband signal of random-interleaved

where, ωc is carrier frequency, *r(t)* is baseband pulse, *<sup>ˆ</sup> <sup>M</sup> T ( ).T <sup>N</sup>* is compressed symbol duration after IFFT and *T* is symbol duration is seconds. The PAPR of the ZCMT precoded SLM based random-interleaved OFDMA uplink signal in (49) with RRC pulse shaping can

Extensive simulations in MATLAB(R) have been carried out to evaluate the performance of the proposed SLM based ZCMT precoded random-interleaved OFDMA uplink system with

(49)

OFDMA after RRC pulse shaping can be written as follows:-

be calculated by equation (43).

**5. Simulation results** 

pulse shaping.

System

Fig. 11. The normalized autocorrelation function

#### **4.7 Selective mapping to further improve ZCMT precoded random interleaved OFDMA system**

Fig.12 shows the block diagram of the proposed SLM based ZCMT precoded randominterleaved OFDMA uplink system. Suppose data stream after S/P conversion is *X*=[*X*0,*X*1,*X*2,…,*XM*-1]*T*, and each data block is multiplied by *V* dissimilar phase sequences, each length equal to *M*, *B(v)* =[*bv,0,bv,1,…,bv,M-1*]*T*,(*v= 1, 2…V),* which results in the altered data blocks. Let us denote the altered data block for the *v*th phase sequence is given by *X(v)* =[*X0bv,0,X1bv,1,…, XN-1bv,M-1*]*T*, where *v* =1,2,3,…*V.* Then, these altered data blocks are passed through the precoder, which transforms this complex vector into new vector of same length *L* that can be written as *Y*=*PX*=[*Y0,Y1,Y2,…, YL-1*]*T*, where *P* is a ZCMT precoder matrix of size *N* =*L*×*L* and *<sup>v</sup> Ym* can be written as follows:-

$$Y\_l^\upsilon = \sum\_{m=0}^{M-1} z\_{m,l} X\_m^\upsilon \qquad l = 0, 1, \dots \\ L - 1 \tag{47}$$

where, *zm,l* means precoding matrix of *mth* row and *lth* column. Equation (47) represents the ZCMT precoded constellations signal. Then the subcarrier mapping of this precoded signal is done in random-interleaved mode. Suppose the *kth* user is assigned to sub-channel *k* then the complex baseband SLM based ZCMT precoded OFDMA uplink signal for *kth* user can be written as follows:-

$$\chi\_n^{(k,\nu)} = \sum\_{l=0}^{M-1} \hat{Y}\_l^{(k,\nu)}.e^{j2\pi \frac{(l\mathfrak{Q}+\nu\_{q,k})}{N}n}, n = 0, 1\ldots N-1\tag{48}$$

*(k,v) <sup>l</sup> Yˆ* is modulated signal on subcarrier *m* for *k*th user.

**4.7 Selective mapping to further improve ZCMT precoded random interleaved OFDMA** 

Fig.12 shows the block diagram of the proposed SLM based ZCMT precoded randominterleaved OFDMA uplink system. Suppose data stream after S/P conversion is *X*=[*X*0,*X*1,*X*2,…,*XM*-1]*T*, and each data block is multiplied by *V* dissimilar phase sequences, each length equal to *M*, *B(v)* =[*bv,0,bv,1,…,bv,M-1*]*T*,(*v= 1, 2…V),* which results in the altered data blocks. Let us denote the altered data block for the *v*th phase sequence is given by *X(v)* =[*X0bv,0,X1bv,1,…, XN-1bv,M-1*]*T*, where *v* =1,2,3,…*V.* Then, these altered data blocks are passed through the precoder, which transforms this complex vector into new vector of same length *L* that can be written as *Y*=*PX*=[*Y0,Y1,Y2,…, YL-1*]*T*, where *P* is a ZCMT precoder matrix of

where, *zm,l* means precoding matrix of *mth* row and *lth* column. Equation (47) represents the ZCMT precoded constellations signal. Then the subcarrier mapping of this precoded signal is done in random-interleaved mode. Suppose the *kth* user is assigned to sub-channel *k* then the complex baseband SLM based ZCMT precoded OFDMA uplink signal for *kth* user can be

(47)

(48)

Fig. 11. The normalized autocorrelation function

size *N* =*L*×*L* and *<sup>v</sup> Ym* can be written as follows:-

*<sup>l</sup> Yˆ* is modulated signal on subcarrier *m* for *k*th user.

**system** 

written as follows:-

*(k,v)*

Fig. 12. Block diagram of SLM based ZCMT precoded random-interleaved OFDMA Uplink System

The precoding based SLM technique needs *V* (dissimilar phase sequences) IFFT operations and the information bits required as side information for each data block is [log2V]. Precoding based SLM technique is applicable for any number of subcarriers and all types of modulation techniques. The PAPR reduction for precoding based SLM technique depends on the number of phase sequences *V* and the output data with lowest PAPR is selected by the transmitter for transmissions. The complex passband signal of random-interleaved OFDMA after RRC pulse shaping can be written as follows:-

$$\mathbf{x}(\mathbf{t}) = \mathbf{e}^{j\,\omega\_{\mathbf{e}}\mathbf{t}} \sum\_{n=0}^{N-1} \mathbf{x}\_{n}^{\{k,v\}} \cdot \mathbf{r}(\mathbf{t} - n\check{T}) \tag{49}$$

where, ωc is carrier frequency, *r(t)* is baseband pulse, *<sup>ˆ</sup> <sup>M</sup> T ( ).T <sup>N</sup>* is compressed symbol duration after IFFT and *T* is symbol duration is seconds. The PAPR of the ZCMT precoded SLM based random-interleaved OFDMA uplink signal in (49) with RRC pulse shaping can be calculated by equation (43).

#### **5. Simulation results**

Extensive simulations in MATLAB(R) have been carried out to evaluate the performance of the proposed SLM based ZCMT precoded random-interleaved OFDMA uplink system with pulse shaping.

On PAPR Reduction Techniques in Mobile WiMAX 211

dB respectively, for the conventional random-interleaved OFDMA uplink systems, WHT precoded random-interleaved OFDMA uplink systems, ZCMT precoded randominterleaved OFDMA uplink systems and SLM based ZCMT precoded random-interleaved

Precoding Based OFDMA System for Mobile-WiMAX (M=16 and N=512 for QPSK)

Conventional-OFDMA WHT-OFDMA ZCMT-OFDMA SLM-ZCMT-OFDMA(V=4)

4 5 6 7 8 9 10 11 12

PAPR0 (dB)

Fig. 13. CCDF Comparison of PAPR of the ZCMT precoded random-interleaved OFDMA uplink system and SLM based ZCMT precoded random-interleaved OFDMA uplink system with the WHT precoded random-interleaved OFDMA uplink system and the conventional

Fig.15 shows CCDF comparison of PAPR for the ZCMT precoded random-interleaved OFDMA uplink system and the SLM based ZCMT precoded random-interleaved OFDMA uplink system with the WHT precoded random-interleaved OFDMA uplink systems and the conventional random-interleaved OFDMA uplink systems. At clip rate of 10-3, with user subcarriers *M*=16 and system subcarriers *N*=512, the PAPR is 9.8 dB, 9.6 dB, 8.7 dB and 6.7 dB respectively, for the conventional random-interleaved OFDMA uplink systems, WHT precoded random-interleaved OFDMA uplink systems, ZCMT precoded randominterleaved OFDMA uplink systems and SLM based ZCMT precoded random-interleaved

Fig.16 shows the SER performance of the ZCMT precoded random-interleaved OFDMA uplink systems; WHT precoded random-interleaved OFDMA uplink systems and the conventional random-interleaved OFDMA uplink systems respectively. The WiMAX Forum recommends using just two out of the six ITU models, which are Pedestrian B and Vehicular A (WiMAX, 2008). So, we use the ITU pedestrian B channel with additive white gaussian noise (AWGN) and MMSE equalization. The parameters for the ITU Pedestrian B channel

random-interleaved OFDMA uplink system respectively, for QPSK modulation.

OFDMA uplink system respectively, using 64-QAM modulation.

model can be found in Table 3 (ITU, 1997).

OFDMA uplink system respectively,, using 16-QAM modulation.

10-3

10-2

Prob(PAPR > PAPR0

)

10-1

100


Table 2. System Parameters

To show PAPR analysis of the proposed system, the data is generated randomly then modulated by QPSK, 16-QAM and 64-QAM respectively. We evaluate the PAPR statistically by using complementary cumulative distribution function (CCDF). The CCDF of the PAPR for ZCMT precoded random interleaved OFDMA uplink signal is used to express the probability of exceeding a given threshold PAPR0 (CCDF = Prob (PAPR > PAPR0)). We compared the simulation results of proposed system with WHT precoded random interleaved OFDMA uplink systems and conventional random interleaved OFDMA uplink systems. To show the PAPR analysis of proposed system with pulse shaping in MATLAB® we considered RRC rolloff factor α = 0.22 with system subcarriers *N=512* and user subcarriers *M=*16. All the simulations have been performed on 105 random data blocks. Simulation parameters that we use are given in the above Table 2.

Fig.13 shows CCDF comparison of PAPR for the ZCMT precoded random-interleaved OFDMA uplink system and the SLM based ZCMT precoded random-interleaved OFDMA uplink system with the WHT precoded random-interleaved OFDMA uplink systems and the conventional random-interleaved OFDMA uplink systems. At clip rate of 10-3, with user subcarriers *M*=16 and system subcarriers *N*=512, the PAPR is 10 dB, 9.2 dB, 7.4 dB and 5.7 dB respectively, for the conventional random-interleaved OFDMA uplink systems, WHT precoded random-interleaved OFDMA uplink systems, ZCMT precoded randominterleaved OFDMA uplink systems and SLM based ZCMT precoded random-interleaved OFDMA uplink system respectively, using QPSK modulation.

Fig.14 shows CCDF comparison of PAPR for the ZCMT precoded random-interleaved OFDMA uplink system and the SLM based ZCMT precoded random-interleaved OFDMA uplink system with the WHT precoded random-interleaved OFDMA uplink systems and the conventional random-interleaved OFDMA uplink systems. At clip rate of 10-3, with user subcarriers *M*=16 and system subcarriers *N*=512, the PAPR is 9.5 dB, 9.3 dB, 8.2 dB and 6.7

Channel Bandwidth 5MHz

Oversampling Factor 8

User Subcarriers 16

System Subcarriers 512

Typical RRC Roll-Off Factor α = 0.22

CCDF Clip Rate 10-3

Simulation parameters that we use are given in the above Table 2.

OFDMA uplink system respectively, using QPSK modulation.

Table 2. System Parameters

Precoding WHT and ZCMT

Modulation QPSK, 16-QAM, 64-QAM

Pulse Shaping Root Raised Cosine (RRC)

Subcarrier Mapping Mode Random Interleaved

To show PAPR analysis of the proposed system, the data is generated randomly then modulated by QPSK, 16-QAM and 64-QAM respectively. We evaluate the PAPR statistically by using complementary cumulative distribution function (CCDF). The CCDF of the PAPR for ZCMT precoded random interleaved OFDMA uplink signal is used to express the probability of exceeding a given threshold PAPR0 (CCDF = Prob (PAPR > PAPR0)). We compared the simulation results of proposed system with WHT precoded random interleaved OFDMA uplink systems and conventional random interleaved OFDMA uplink systems. To show the PAPR analysis of proposed system with pulse shaping in MATLAB® we considered RRC rolloff factor α = 0.22 with system subcarriers *N=512* and user subcarriers *M=*16. All the simulations have been performed on 105 random data blocks.

Fig.13 shows CCDF comparison of PAPR for the ZCMT precoded random-interleaved OFDMA uplink system and the SLM based ZCMT precoded random-interleaved OFDMA uplink system with the WHT precoded random-interleaved OFDMA uplink systems and the conventional random-interleaved OFDMA uplink systems. At clip rate of 10-3, with user subcarriers *M*=16 and system subcarriers *N*=512, the PAPR is 10 dB, 9.2 dB, 7.4 dB and 5.7 dB respectively, for the conventional random-interleaved OFDMA uplink systems, WHT precoded random-interleaved OFDMA uplink systems, ZCMT precoded randominterleaved OFDMA uplink systems and SLM based ZCMT precoded random-interleaved

Fig.14 shows CCDF comparison of PAPR for the ZCMT precoded random-interleaved OFDMA uplink system and the SLM based ZCMT precoded random-interleaved OFDMA uplink system with the WHT precoded random-interleaved OFDMA uplink systems and the conventional random-interleaved OFDMA uplink systems. At clip rate of 10-3, with user subcarriers *M*=16 and system subcarriers *N*=512, the PAPR is 9.5 dB, 9.3 dB, 8.2 dB and 6.7 dB respectively, for the conventional random-interleaved OFDMA uplink systems, WHT precoded random-interleaved OFDMA uplink systems, ZCMT precoded randominterleaved OFDMA uplink systems and SLM based ZCMT precoded random-interleaved OFDMA uplink system respectively,, using 16-QAM modulation.

Fig. 13. CCDF Comparison of PAPR of the ZCMT precoded random-interleaved OFDMA uplink system and SLM based ZCMT precoded random-interleaved OFDMA uplink system with the WHT precoded random-interleaved OFDMA uplink system and the conventional random-interleaved OFDMA uplink system respectively, for QPSK modulation.

Fig.15 shows CCDF comparison of PAPR for the ZCMT precoded random-interleaved OFDMA uplink system and the SLM based ZCMT precoded random-interleaved OFDMA uplink system with the WHT precoded random-interleaved OFDMA uplink systems and the conventional random-interleaved OFDMA uplink systems. At clip rate of 10-3, with user subcarriers *M*=16 and system subcarriers *N*=512, the PAPR is 9.8 dB, 9.6 dB, 8.7 dB and 6.7 dB respectively, for the conventional random-interleaved OFDMA uplink systems, WHT precoded random-interleaved OFDMA uplink systems, ZCMT precoded randominterleaved OFDMA uplink systems and SLM based ZCMT precoded random-interleaved OFDMA uplink system respectively, using 64-QAM modulation.

Fig.16 shows the SER performance of the ZCMT precoded random-interleaved OFDMA uplink systems; WHT precoded random-interleaved OFDMA uplink systems and the conventional random-interleaved OFDMA uplink systems respectively. The WiMAX Forum recommends using just two out of the six ITU models, which are Pedestrian B and Vehicular A (WiMAX, 2008). So, we use the ITU pedestrian B channel with additive white gaussian noise (AWGN) and MMSE equalization. The parameters for the ITU Pedestrian B channel model can be found in Table 3 (ITU, 1997).

On PAPR Reduction Techniques in Mobile WiMAX 213

SER vs. SNR Using QPSK modulation(ITU Pedestrian B Channel,MMSE)

0 5 10 15 20 25 30

**WHT Precoded Random-Interleaved OFDMA**

**Conventional Random-Interleaved OFDMA**

> **ZCMT Precoded Random-Interleaved OFDMA**

SNR(dB)

It is concluded from Fig.16 that the ZCMT precoded random-interleaved OFDMA uplink systems provides approximately same performance as that of the WHT precoded randominterleaved OFDMA systems but a significant SER performance improvement is seen over the conventional random interleaved OFDMA uplink systems for the sub-band 0 using

Table 4 summarizes the PAPR of random-interleaved OFDMA uplink systems, WHT random-interleaved OFDMA uplink systems, ZCMT precoded random-interleaved OFDMA uplink systems and SLM based ZCMT precoded random-interleaved OFDMA uplink systems respectively, using QPSK, 16-QAM and 64-QAM. Table 4 concludes that, the ZCMT precoded random-interleaved OFDMA uplink system and the SLM based ZCMT precoded random-interleaved OFDMA uplink system has lower PAPR than the WHT precoded random-interleaved OFDMA uplink systems and conventional random-interleaved

**Tap No. 1 2 3 4 5 6** 

**Delay (ns)** 0 200 800 1200 2300 3700

**Power (dB)** 0.0 -0.9 -4.9 -8.0 -7.8 -23.9

**Spectrum** Classic Classic Classic Classic Classic Classic

Fig. 16. SER vs. SNR Comparison of the ZCMT precoded Random-Interleaved OFDMA uplink system, the WHT precoded Random- Interleaved OFDMA uplink system and the conventional Random-Interleaved OFDMA uplink system, for sub-band 0 with QPSK

10-4

modulation.

QPSK modulation.

OFDMA uplink systems.

**Relative** 

**Average** 

**Doppler** 

Table 3. ITU Pedestrian B channel Parameters

**ITU** 

**Pedestrian B Channel** 

10-3

10-2

SER

10-1

100

Fig. 14. CCDF Comparison of PAPR of the ZCMT precoded-random interleaved OFDMA uplink system and SLM based ZCMT precoded random-interleaved OFDMA uplink system with the WHT precoded random-interleaved OFDMA uplink system and the conventional random-interleaved OFDMA uplink system respectively, for 16-QAM modulation.

Fig. 15. CCDF Comparison of PAPR of the ZCMT precoded random-interleaved OFDMA uplink system and SLM based ZCMT precoded random-interleaved OFDMA uplink system with the WHT precoded random-interleaved OFDMA uplink system and the conventional random-interleaved OFDMA uplink system respectively, for 64-QAM modulation.

Precoding Based OFDMA System for Mobile-WiMAX (M=16 and N=512 for 16-QAM)

Conventional-OFDMA WHT-OFDMA ZCMT-OFDMA SLM-ZCMT-OFDMA(V=4)

Conventional-OFDMA WHT-OFDMA ZCMT-OFDMA SLM-ZCMT-OFDMA(V=4)

<sup>4</sup> <sup>5</sup> <sup>6</sup> <sup>7</sup> <sup>8</sup> <sup>9</sup> <sup>10</sup> <sup>11</sup> <sup>12</sup> 10-3

PAPR0 (dB)

Precoding Based OFDMA System for Mobile-WiMAX (M=16 and N=512 for 64-QAM)

4 5 6 7 8 9 10 11 12

PAPR0 (dB)

Fig. 15. CCDF Comparison of PAPR of the ZCMT precoded random-interleaved OFDMA uplink system and SLM based ZCMT precoded random-interleaved OFDMA uplink system with the WHT precoded random-interleaved OFDMA uplink system and the conventional

random-interleaved OFDMA uplink system respectively, for 64-QAM modulation.

Fig. 14. CCDF Comparison of PAPR of the ZCMT precoded-random interleaved OFDMA uplink system and SLM based ZCMT precoded random-interleaved OFDMA uplink system with the WHT precoded random-interleaved OFDMA uplink system and the conventional

random-interleaved OFDMA uplink system respectively, for 16-QAM modulation.

10-2

10-3

10-2

Prob(PAPR > PAPR0

)

10-1

100

Prob(PAPR > PAPR0

)

10-1

100

Fig. 16. SER vs. SNR Comparison of the ZCMT precoded Random-Interleaved OFDMA uplink system, the WHT precoded Random- Interleaved OFDMA uplink system and the conventional Random-Interleaved OFDMA uplink system, for sub-band 0 with QPSK modulation.

It is concluded from Fig.16 that the ZCMT precoded random-interleaved OFDMA uplink systems provides approximately same performance as that of the WHT precoded randominterleaved OFDMA systems but a significant SER performance improvement is seen over the conventional random interleaved OFDMA uplink systems for the sub-band 0 using QPSK modulation.

Table 4 summarizes the PAPR of random-interleaved OFDMA uplink systems, WHT random-interleaved OFDMA uplink systems, ZCMT precoded random-interleaved OFDMA uplink systems and SLM based ZCMT precoded random-interleaved OFDMA uplink systems respectively, using QPSK, 16-QAM and 64-QAM. Table 4 concludes that, the ZCMT precoded random-interleaved OFDMA uplink system and the SLM based ZCMT precoded random-interleaved OFDMA uplink system has lower PAPR than the WHT precoded random-interleaved OFDMA uplink systems and conventional random-interleaved OFDMA uplink systems.


Table 3. ITU Pedestrian B channel Parameters

On PAPR Reduction Techniques in Mobile WiMAX 215

Cho, Y. S.; Kim, J.; Yang, W. Y. & Kang, C. G. (2011). MIMO-OFDM Wireless

Chu, D.C. (1972). Polyphase codes with good periodic correlation properties. *IEEE Trans.* 

Cimini, L. J. & Sollenberger, N. R. (2000). Peak-to-Average Power Ratio Reduction of an

Han, S. H. & Lee, J. H. (2004). PAPR Reduction of OFDM Signals Using a Reduced

Han, S. H. & Lee, J. H. (2005). An overview of peak-to-average power Ratio reduction a

ITU, Document. (1997). Rec. ITU-R M.1225-Guidelines for Evaluation of Radio Transmission

Jayalath, A. & Tellambura, C. (2000). Adaptive PTS Approach for Reduction of Peak-to-

Jeoti, V. & Baig, I. (2009). A Novel Zadoff-Chu Precoder Based SLM Technique for PAPR

Knopp, R. & Humblet, P. A. (1995). Information Capacity and Power Control in Single-Cell Multiuser Communications. *IEEE Int. Conf. Communication*, pp. 331-335, 1995. Kou, Y.; Lu, W. & Antoniou, A. (2007). A new peak-to- average power-ratio reduction

Li, X. & Cimini, L.J. (1998) Effects of clipping and filtering on the performance of OFDM.

Lim, D. W.; No, J. S.; Lim, C. W. & Chung, H.(2005). A new SLM OFDM scheme with low

Min, Y. K. & Jeoti, V. (2007). A Novel Signal Independent Technique for PAPR Reduction in

Mourelo, J. T.(1999). Peak to Average Power Ratio Reduction for Multicarrier Modulation.

Müller, S. H. & Huber, J. B. (1997). A Novel Peak Power Reduction Scheme for OFDM,"

Nee, V. & Wild, A. (1998) Reducing the peak-to-average power ratio of OFDM. IEEE

Nikookar, H. & Lidsheim, K. S. (2002). Random phase updating algorithm for OFDM

transmission with low PAPR. *IEEE Trans. Broadcasting*, vol. 48, no. 2, pp. 123–128,

*Proc. IEEE PIMRC '97*, Helsinki, Finland, pp. 1090–94, Sept. 1997.

*Conference on Antennas, Propagation and Systems*, Johor, Malaysia, 2009. Jiang, T.; Yao, W.; Guo, P.; Song, Y. & Qu, D. (2006). Two novel nonlinear Companding

systems. *IEEE Trans. Broadcasting*, vol. 52, no. 2, pp. 268–273, 2006.

*Communications*, vol. 6, no. 5, pp. 1823–1832, 2007.

OFDM Systems. *IEEE-ICSCN*, pp. 308-311, 2007.

*PhD thesis, University of Stanford, Stanford*, 1999.

VTC'98, vol.3, pp. 18–21, May 1998.

IEEE Commun. Letter, 2(20), 131–133, 1998.

*Inform. Theory*, vol. IT-18, pp. 531-532, 1972.

Technologies for IMT-2000, ITU-R, 1997.

470-82561-7.

890, 2004.

2000.

96, 2005.

2002.

pp. 86–88, Mar. 2000.

issue.2, pp: 56-65, 2005.

Communications with Matlab. John Wiley & Sons, Inc. (Asia) Pte Ltd. ISBN:978-0-

OFDM Signal Using Partial Transmit Sequences. *IEEE Commun. Lett.*, vol. 4, no. 3,

Complexity PTS Technique. *Signal Processing Letters, IEEE*, vol.11, issue.11, pp: 887-

techniques for multicarrier transmission. IEEE Wireless communication, vol.12,

Average Power Ratio of OFDM Signal," *Elect. Lett.*, vol. 36, no. 14, pp. 1226–28, July

Reduction in OFDM Systems. invited paper, *Proceedings of 2009 IEEE International* 

schemes with iterative receiver to reduce PAPR in multicarrier modulation

algorithm for OFDM systems via constellation extension. *IEEE Trans. Wireless* 

complexity for PAPR reduction. *IEEE Signal Processing Letters*, vol. 12, no. 2, pp. 93–


Table 4. At CCDF of 10-3, The PAPR Comparisons of the Conventional Random-Interleaved OFDMA uplink, WHT Random-Interleaved OFDMA uplink, ZCMT Random-Interleaved OFDMA uplink and SLM based ZCMT Random-Interleaved OFDMA uplink respectively, for users subcarriers (*M*=16) and system subcarriers (*N*=512)

#### **6. Conclusion**

In this chapter, we present a brief overview of the mobile WiMAX and typical PAPR reduction techniques available in the literature. We also introduce two precoding based systems: ZCMT precoded random-interleaved OFDMA uplink system and SLM based ZCMT precoded random-interleaved OFDMA uplink system, for PAPR reduction in mobile WiMAX systems. Computer simulation shows that the PAPR of the both proposed systems have less PAPR than the WHT precoded random-interleaved OFDMA uplink systems and conventional random-interleaved OFDMA uplink systems. These systems are efficient, signal independent, distortionless and do not require any complex optimizations. Additionally, these systems also take the advantage of the frequency variations of the communication channel and can also offer substantial performance gain in fading multipath channels. Thus, it is concluded that the both proposed uplink systems are more favourable than the WHT precoded random-interleaved OFDMA uplink systems and conventional random-interleaved OFDMA uplink systems for the mobile WiMAX systems.

#### **7. References**


Conventional Random-Interleaved OFDMA 10 dB 9.5 dB 9.8 dB WHT Random-Interleaved OFDMA 9.2 dB 9.3 dB 9.6 dB ZCMT Random-Interleaved OFDMA 7.4 dB 8.2 dB 8.7 dB SLM-ZCMT Random-Interleaved OFDMA 5.7 dB 6.7 dB 6.7 dB

Table 4. At CCDF of 10-3, The PAPR Comparisons of the Conventional Random-Interleaved OFDMA uplink, WHT Random-Interleaved OFDMA uplink, ZCMT Random-Interleaved OFDMA uplink and SLM based ZCMT Random-Interleaved OFDMA uplink respectively,

In this chapter, we present a brief overview of the mobile WiMAX and typical PAPR reduction techniques available in the literature. We also introduce two precoding based systems: ZCMT precoded random-interleaved OFDMA uplink system and SLM based ZCMT precoded random-interleaved OFDMA uplink system, for PAPR reduction in mobile WiMAX systems. Computer simulation shows that the PAPR of the both proposed systems have less PAPR than the WHT precoded random-interleaved OFDMA uplink systems and conventional random-interleaved OFDMA uplink systems. These systems are efficient, signal independent, distortionless and do not require any complex optimizations. Additionally, these systems also take the advantage of the frequency variations of the communication channel and can also offer substantial performance gain in fading multipath channels. Thus, it is concluded that the both proposed uplink systems are more favourable than the WHT precoded random-interleaved OFDMA uplink systems and conventional

Baig, I. & Jeoti, V. (2010). DCT Precoded SLM Technique for PAPR Reduction in OFDM

Baig, I. & Jeoti, V. (2010). Novel Precoding Based PAPR Reduction Techniques for Localized-

Baig, I. & Jeoti, V. (2010). PAPR Reduction in OFDM Systems: Zadoff-Chu Matrix Transform

*and Computer Engineering (JTEC) Malaysia*, vol.2 no.1, pp 49-58, 2010. Baig, I. & Jeoti, V. (2010). PAPR Analysis of DHT-Precoded OFDM System for M-QAM. *The* 

*Intelligence, Communication Systems and Networks, Liverpool, UK,* 2010.

Systems," *The 3rd International Conference on Intelligent and Advanced Systems, Kuala* 

OFDMA Uplink System of LTE-Advanced. *Journal of Telecommunication, Electronic* 

*3rd International Conference on Intelligent and Advanced Systems,* Kuala Lumpur,

Based Pre/Post-Coding Techniques. *2nd International Conference on Computational* 

random-interleaved OFDMA uplink systems for the mobile WiMAX systems.

for users subcarriers (*M*=16) and system subcarriers (*N*=512)

**PAPR QPSK 16-QAM 64-QAM** 

**Transmission Scheme** 

**6. Conclusion** 

**7. References** 

*Lumpur, Malaysia*, 2010.

Malaysia*,* 2010.


**11** 

**Peak-to-Average Power Ratio** 

Pooria Varahram and Borhanuddin Mohd Ali

Broadband wireless is a technology that provides connection over the air at high speeds. Orthogonal frequency division multiplexing (OFDM) system has generally been adopted in recent mobile communication systems because of its high spectral efficiency and robustness against intersymbol interference (ISI). However, due to the nature of inverse fast Fourier transform (IFFT) in which the constructive and destructive behaviour could create high peak signal in constructive behaviour while the average can become zero at destructive behaviour, OFDM signals generally become prone to high peak-to-average power ratio (PAPR) problem. In this chapter, we focus on some of the techniques to overcome the PAPR

The other issue in wireless broadband is how to maximize the power efficiency of the power amplifier. This can be resolved by applying digital predistortion to the power amplifier (PA) (Varahram, et al. 2009). High PAPR signal when transmitted through a nonlinear PA creates spectral broadening and increase the dynamic range requirement of the digital to analog converter (DAC). This results in an increase in the cost of the system and a reduction in efficiency. To address this problem, many techniques for reducing PAPR have been proposed. Some of the most important techniques are clipping (Kwon, et al. 2009), windowing (Van Nee and De Wild, 1998), envelope scaling (Foomooljareon and Fernando, 2002), random phase updating (Nikookar and Lidsheim, 2002), peak reduction carrier (Tan and Wassell, 2003), companding (Hao and Liaw, 2008), coding (Wilkison and Jones, 1995), selected mapping (SLM) (Bauml, et al. 1996), partial transmit sequence (PTS) (Muller and Huber, 1997), DSI-PTS (Varahram et al. 2010), interleaving (Jayalath and Tellambura, 2000), active constellation extension (ACE) (Krongold, et al. 2003), tone injection and tone reservation (Tellado, 2000), dummy signal insertion (DSI) (Ryu, et al. 2004), addition of

Clipping is the simplest technique for PAPR reduction, where the signal at the transmitter is clipped to a desired level without modifying the phase information. In windowing a peak of the signal is multiplied with a part of the frame. This frame can be

problem (Krongold and Jones, 2003; Bauml, et al. 1996).

Guassian signals (Al-Azoo et al. 2008) and etc (Qian, 2005).

**1. Introduction** 

**Reduction in Orthogonal** 

**Frequency Division** 

*Universiti Putra Malaysia,* 

*Malaysia* 

**Multiplexing Systems** 


### **Peak-to-Average Power Ratio Reduction in Orthogonal Frequency Division Multiplexing Systems**

Pooria Varahram and Borhanuddin Mohd Ali *Universiti Putra Malaysia, Malaysia* 

#### **1. Introduction**

216 Advanced Transmission Techniques in WiMAX

Popovic´, B. M. (1997). Spreading sequences for multi-carrier CDMA systems. *in IEE* 

Slimane, S. B. (2007). Reducing the peak-to-average power ratio of OFDM signals through precoding. *IEEE Trans. Vehicular Technology*, vol.56, no. 2, pp. 686–695, Mar. 2007. Tasi, Y.; Zhang, G. & Wang, X. (2006). Orthogonal Polyphase Codes for Constant Envelope

Tellambura, C. (1997). Upper bound on peak factor of N-multiple carriers. Electronics

Tellambura, C. (2001). Improved Phase Factor Computation for the PAR Reduction of an OFDM Signal Using PTS. *IEEE Commun. Lett.*, vol. 5, no. 4, pp. 135–37, Apr. 2001. Thompson, S. C.; Ahmed, A. U.; Proakis, J. G.; Zeidler, J. R. & Geile, M. J. (2008).Constant envelope OFDM. *IEEE Trans. Communications*, vol. 56, pp. 1300-1312, 2008. Tse, D. (1997). Optimal Power Allocation over Parallel Gaussian Broadcast Channels. *Proceedings of International Symposium on Information, Ulm Germany*, pp. 27, 1997. Wang, H. & Chen, B. (2004). Asymptotic distributions and peak power analysis for uplink

Wang, L. & Tellambura, C. (2005). A Simplified Clipping and Filtering Technique for PAR

WiMAX, Forum. (2008). WiMAX System Evaluation Methodology. Version 2.1, July 2008. Yoo, S.; Yoon, S.; Kim, S.Y. & Song, I. (2006). A novel PAPR reduction scheme for OFDM

OFDMA signals. *in Proc. IEEE Acoustics, Speech, and Signal Processing Conference*,

Reduction in OFDM Systems. *Signal Processing Letters, IEEE*, vol.12, no.6, pp. 453-

systems: Selective mapping of partial tones (SMOPT). *IEEE Trans. Consumer* 

OFDM-CDMA System. *IEEE, WCNC*, pp.1396 – 1401, 2006.

*London*, pp. 8/1–8/6, 1997.

vol.4, pp.1085-1088, 2004.

456, 2005.

Letters, vol.33, pp.1608-1609, Sept.1997.

*Electronics*, vol. 52, no. 1, pp.40–43, 2006.

*Colloquium CDMA Techniques and Applications for Third Generation Mobile Systems,* 

Broadband wireless is a technology that provides connection over the air at high speeds. Orthogonal frequency division multiplexing (OFDM) system has generally been adopted in recent mobile communication systems because of its high spectral efficiency and robustness against intersymbol interference (ISI). However, due to the nature of inverse fast Fourier transform (IFFT) in which the constructive and destructive behaviour could create high peak signal in constructive behaviour while the average can become zero at destructive behaviour, OFDM signals generally become prone to high peak-to-average power ratio (PAPR) problem. In this chapter, we focus on some of the techniques to overcome the PAPR problem (Krongold and Jones, 2003; Bauml, et al. 1996).

The other issue in wireless broadband is how to maximize the power efficiency of the power amplifier. This can be resolved by applying digital predistortion to the power amplifier (PA) (Varahram, et al. 2009). High PAPR signal when transmitted through a nonlinear PA creates spectral broadening and increase the dynamic range requirement of the digital to analog converter (DAC). This results in an increase in the cost of the system and a reduction in efficiency. To address this problem, many techniques for reducing PAPR have been proposed. Some of the most important techniques are clipping (Kwon, et al. 2009), windowing (Van Nee and De Wild, 1998), envelope scaling (Foomooljareon and Fernando, 2002), random phase updating (Nikookar and Lidsheim, 2002), peak reduction carrier (Tan and Wassell, 2003), companding (Hao and Liaw, 2008), coding (Wilkison and Jones, 1995), selected mapping (SLM) (Bauml, et al. 1996), partial transmit sequence (PTS) (Muller and Huber, 1997), DSI-PTS (Varahram et al. 2010), interleaving (Jayalath and Tellambura, 2000), active constellation extension (ACE) (Krongold, et al. 2003), tone injection and tone reservation (Tellado, 2000), dummy signal insertion (DSI) (Ryu, et al. 2004), addition of Guassian signals (Al-Azoo et al. 2008) and etc (Qian, 2005).

Clipping is the simplest technique for PAPR reduction, where the signal at the transmitter is clipped to a desired level without modifying the phase information. In windowing a peak of the signal is multiplied with a part of the frame. This frame can be

Peak-to-Average Power Ratio Reduction in

given by (Han and Lee, 2005):

the Gaussian *Q* function defined as:

**3. PAPR reduction techniques** 

**3.1 Conventional SLM (C-SLM)** 

ratio (*E*b/*N*o) is analyzed.

Orthogonal Frequency Division Multiplexing Systems 219

given level. Knowing the amplitude distribution of the OFDM output signals, it is easy to compute the probability that the instantaneous amplitude will lie above a given threshold and the same goes for power. This is performed by calculating the complementary

Here the effect of additive white Gaussian noise (AWGN) on OFDM performance is studied. As OFDM systems use standard digital modulation formats to modulate the subcarriers, PSK and QAM are usually used due to their excellent error resilient properties. The most important block in OFDM is IFFT. IFFT changes the distribution of the signal without altering its average power. The BER or bit error probability *Pbe* in an AWGN channel is

*4( M 1) 3k E P Q. k M (M 1) N*

where *M* is the modulation order, *k*= log2(*M*) is the number of bits per symbol, and *Q*(.) is

*<sup>y</sup> Q(y) erfc( ) <sup>2</sup>* (6)

In this chapter the performance of BER versus energy per bit to noise power spectral density

In this section, some of the most important PAPR reduction techniques such as Selected Mapping (SLM), Partial Transmit Sequence (PTS) and Enhanced PTS EPTS) are presented.

In Conventional SLM (C-SLM) method, OFDM signal is first converted from serial to parallel by means of serial-to-parallel converter. The parallel OFDM signal is then multiplied by several phase sequences that are created offline and stored in a matrix. A copy of the OFDM signal is multiplied with a random vector of phase sequence matrix. For each subblock IFFT is performed and its PAPR is calculated to look for the minimum one. The OFDM signal having minimum PAPR is then selected and be transmitted. The main drawbacks of this technique are the high complexity due to the high number of subblocks and the need to send side information which result in data rate and transmission efficiency degradation, respectively. In Fig. 1, the number of candidate signal or subblocks is given by

The other drawback of this method is that by increasing *U*, higher number of IFFT blocks are required which increase the complexity significantly. Hence, a method with low

*U*, hence *<sup>2</sup> log U* number of bits is required to be sent as side information.

complexity and high PAPR performance is required.

*CCDF Pr(PAPR PAPR ) <sup>0</sup>* (4)

*o*

(5)

cumulative distribution function (CCDF) for different PAPR values as follows:

 *<sup>b</sup> be,MQAM*

in Gaussian shape, cosine, Kaiser or Hanning window, respectively. In companding method the OFDM signal is companded before digital to analog conversion. The OFDM signal after IFFT is first companded and quantized and then transmitted through the channel after digital to analog conversion. The receiver first converts the signal into digital format and then expands it. The companding method has application in speech processing where high peaks occur infrequently. In PTS, by partitioning the input signal and applying several IFFT, the optimum phase sequence with lowest PAPR will be selected before being transmitted. This technique results in high complexity. In SLM, a copy of input signal is used to choose the minimum PAPR among the multiple signals. We can conclude that there is always a trade-off in choosing a particular PAPR technique. The trade-off comes in the form of complexity, power amplifier output distortion, cost, side information, PAPR reduction, Bit Error Rate (BER) performance, spectrum efficiency and data rate loss.

#### **2. OFDM signal**

In OFDM systems, first a specific number of input data samples are modulated (e.g. PSK or QAM), and by IFFT technique the input samples become orthogonal and will be converted to time domain at the transmitter side. The IFFT is applied to produce orthogonal data subcarriers. In theory, IFFT combines all the input signals (superposition process) to produce each element (signal) of the output OFDM symbol. The time domain complex baseband OFDM signal can be represented as (Han and Lee, 2005):

$$\mathbf{x}\_n = \frac{1}{\sqrt{N}} \sum\_{k=0}^{N-1} X\_k e^{j2\pi \frac{\mathbf{n}}{N}k}, \quad n = 0, 1, 2, \dots, N-1 \tag{1}$$

where *xn* is the n-th signal component in OFDM output symbol, *Xk* is the *k*-th data modulated symbol in OFDM frequency domain, and *N* is the number of subcarrier.

The PAPR of the transmitted OFDM signal can be given by (Cimini and Sollenberger, 2000):

$$PAPR(dB) = \frac{\max\left[\left|\mathbf{x}\_n\right|^2\right]}{E\left[\left|\mathbf{x}\_n\right|^2\right]}\tag{2}$$

where *E .* is the expectation value operator. The theoretical maximum of PAPR for *N* number of subcarriers is as follows:

$$PAPR\_{\text{max}} = 10\log(\text{N})\text{ dB} \tag{3}$$

PAPR is a random variable since it is a function of the input data, while the input data is a random variable. Therefore PAPR can be analyzed by using level crossing rate theorem which calculates the mean number of times that the envelope of a stationary signal crosses a

in Gaussian shape, cosine, Kaiser or Hanning window, respectively. In companding method the OFDM signal is companded before digital to analog conversion. The OFDM signal after IFFT is first companded and quantized and then transmitted through the channel after digital to analog conversion. The receiver first converts the signal into digital format and then expands it. The companding method has application in speech processing where high peaks occur infrequently. In PTS, by partitioning the input signal and applying several IFFT, the optimum phase sequence with lowest PAPR will be selected before being transmitted. This technique results in high complexity. In SLM, a copy of input signal is used to choose the minimum PAPR among the multiple signals. We can conclude that there is always a trade-off in choosing a particular PAPR technique. The trade-off comes in the form of complexity, power amplifier output distortion, cost, side information, PAPR reduction, Bit Error Rate (BER) performance,

In OFDM systems, first a specific number of input data samples are modulated (e.g. PSK or QAM), and by IFFT technique the input samples become orthogonal and will be converted to time domain at the transmitter side. The IFFT is applied to produce orthogonal data subcarriers. In theory, IFFT combines all the input signals (superposition process) to produce each element (signal) of the output OFDM symbol. The time domain complex

*<sup>1</sup> x X e , n 0,1,2,.......,N 1 <sup>N</sup>*

where *xn* is the n-th signal component in OFDM output symbol, *Xk* is the *k*-th data

The PAPR of the transmitted OFDM signal can be given by (Cimini and Sollenberger,

*PAPR(dB) <sup>2</sup>*

where *E .* is the expectation value operator. The theoretical maximum of PAPR for *N*

PAPR is a random variable since it is a function of the input data, while the input data is a random variable. Therefore PAPR can be analyzed by using level crossing rate theorem which calculates the mean number of times that the envelope of a stationary signal crosses a

(1)

*2*

*PAPR 10log(N ) max dB* (3)

(2)

*max xn*

*E xn* 

baseband OFDM signal can be represented as (Han and Lee, 2005):

*N 1 <sup>n</sup> j2 k <sup>N</sup> <sup>k</sup> <sup>n</sup> k 0*

modulated symbol in OFDM frequency domain, and *N* is the number of subcarrier.

spectrum efficiency and data rate loss.

number of subcarriers is as follows:

**2. OFDM signal** 

2000):

given level. Knowing the amplitude distribution of the OFDM output signals, it is easy to compute the probability that the instantaneous amplitude will lie above a given threshold and the same goes for power. This is performed by calculating the complementary cumulative distribution function (CCDF) for different PAPR values as follows:

$$\text{CCDF} = \Pr(\text{PAPR} > \text{PAPR}\_0) \tag{4}$$

Here the effect of additive white Gaussian noise (AWGN) on OFDM performance is studied. As OFDM systems use standard digital modulation formats to modulate the subcarriers, PSK and QAM are usually used due to their excellent error resilient properties. The most important block in OFDM is IFFT. IFFT changes the distribution of the signal without altering its average power. The BER or bit error probability *Pbe* in an AWGN channel is given by (Han and Lee, 2005):

$$P\_{he,MQAM} \approx \frac{4(\sqrt{M}-1)}{k\sqrt{M}}Q\left(\sqrt{\frac{3k}{(M-1)}}, \frac{E\_b}{N\_o}\right) \tag{5}$$

where *M* is the modulation order, *k*= log2(*M*) is the number of bits per symbol, and *Q*(.) is the Gaussian *Q* function defined as:

$$Q(y) = \text{erfc}(\frac{y}{\sqrt{2}}) \tag{6}$$

In this chapter the performance of BER versus energy per bit to noise power spectral density ratio (*E*b/*N*o) is analyzed.

#### **3. PAPR reduction techniques**

In this section, some of the most important PAPR reduction techniques such as Selected Mapping (SLM), Partial Transmit Sequence (PTS) and Enhanced PTS EPTS) are presented.

#### **3.1 Conventional SLM (C-SLM)**

In Conventional SLM (C-SLM) method, OFDM signal is first converted from serial to parallel by means of serial-to-parallel converter. The parallel OFDM signal is then multiplied by several phase sequences that are created offline and stored in a matrix. A copy of the OFDM signal is multiplied with a random vector of phase sequence matrix. For each subblock IFFT is performed and its PAPR is calculated to look for the minimum one. The OFDM signal having minimum PAPR is then selected and be transmitted. The main drawbacks of this technique are the high complexity due to the high number of subblocks and the need to send side information which result in data rate and transmission efficiency degradation, respectively. In Fig. 1, the number of candidate signal or subblocks is given by *U*, hence *<sup>2</sup> log U* number of bits is required to be sent as side information.

The other drawback of this method is that by increasing *U*, higher number of IFFT blocks are required which increase the complexity significantly. Hence, a method with low complexity and high PAPR performance is required.

Peak-to-Average Power Ratio Reduction in

series

Fig. 3.

following condition:

obtaining the optimum *b*

For finding the optimum *b*

**3.3 Enhanced PTS (EPTS)** 

Orthogonal Frequency Division Multiplexing Systems 221

Fig. 2. Block diagram of the C-PTS scheme with Digital predistortion and power amplifier in

*b argmin( max b x )*

where *V* is the number of subblocks partitioning and *F* is the oversampling factor. After

since one phase factor can remain fixed, *b1=*1. Hence to find the optimum phase factor, *WV-1*

In order to decrease the complexity of C-PTS, a new phase sequence is generated. The block diagram of the enhanced partial transmit sequence (EPTS) scheme is shown in

This new phase sequence is based on the generation of *N* random values of {1 -1 j –j} if the

*V v v 0 k NF 1 v 1*

(10)

, we should perform exhaustive search for (*V-1*) phase factors

which satisfies the

This process is performed by choosing the optimization parameter *b*

, the signal is transmitted.

iteration should be performed, where *W* is the number of allowed phase factors.

allowed phase factors is *W*=4. The phase sequence matrix can be given by:

Fig. 1. The block diagram of the C-SLM method**.**

#### **3.2 Conventional PTS (C-PTS)**

To analyze C-PTS let X denotes random input signal in frequency domain with length *N*. *X* is partitioned into *V* disjoint subblocks *Xv=[Xv,0,Xv,1,…,Xv,N-1]T*, *v=*1,2,…,*V* such that *V v v 1 X X* and then these subblocks are combined to minimize the PAPR in time domain. The Sbblock partitioning is based on interleaving in which the computational complexity is less compared to adjacent and pseudo-random, however it gives the worst PAPR

performance among them (Han and Lee, 2005). By applying the phase rotation factor *<sup>v</sup> <sup>j</sup> <sup>v</sup> b e ,v 1,2,...,V* to the IFFT of the *v*th subblock Xv,

$$\mathbf{x}'(b) = \sum\_{v=1}^{V} b\_v \mathbf{x}\_v \tag{7}$$

where *<sup>T</sup> 0 1 NF 1 x (b) [x (b),x (b),...x (b)]* . The objective is to find the optimum signal *x(b)* with the lowest PAPR.

Both b and x can be shown in matrix forms as follows:

the time domain signal after combining is obtained as:

$$b = \begin{bmatrix} b\_1, & b\_1, \dots, b\_1 \\ \vdots & \vdots & \vdots \\ b\_{V'}, b\_{V'}, \dots, b\_{V} \end{bmatrix}\_{V \times N} \tag{8}$$

$$\mathbf{x} = \begin{vmatrix} \mathbf{x}\_{1,0}, \mathbf{x}\_{1,1}, \dots, \mathbf{x}\_{1,NF-1} \\ \vdots & \vdots & \vdots \\ \mathbf{x}\_{V,0}, \mathbf{x}\_{V,1}, \dots, \mathbf{x}\_{V,NF-1} \end{vmatrix}\_{V \times NF} \tag{9}$$

Fig. 2 shows the block diagram of C-PTS. It should be noted that all the elements of each row of matrix b are of the same values and this is in accordance with the C-PTS method. In order to obtain exact PAPR calculation, at least four times oversampling is necessary (Han and Lee, 2005).

To analyze C-PTS let X denotes random input signal in frequency domain with length *N*. *X* is partitioned into *V* disjoint subblocks *Xv=[Xv,0,Xv,1,…,Xv,N-1]T*, *v=*1,2,…,*V* such that

and then these subblocks are combined to minimize the PAPR in time domain.

The Sbblock partitioning is based on interleaving in which the computational complexity is less compared to adjacent and pseudo-random, however it gives the worst PAPR

*V*

*11 1*

 

*b , b ,...., b*

*1,0 1,1 1,NF 1*

 

Fig. 2 shows the block diagram of C-PTS. It should be noted that all the elements of each row of matrix b are of the same values and this is in accordance with the C-PTS method. In order to obtain exact PAPR calculation, at least four times oversampling is necessary (Han

*x ,x ,...,x*

*x ,x ,...,x*

*VV V V N*

*V ,0 V ,1 V ,NF 1 V NF*

*b , b ,..., b*

*v v v 1 x(b) b x* 

*0 1 NF 1 x (b) [x (b),x (b),...x (b)]* . The objective is to find the optimum signal *x(b)* with

*b*

*x*

*<sup>v</sup> b e ,v 1,2,...,V* to the IFFT of the *v*th subblock Xv,

(7)

(8)

(9)

Fig. 1. The block diagram of the C-SLM method**.**

performance among them (Han and Lee, 2005).

the time domain signal after combining is obtained as:

Both b and x can be shown in matrix forms as follows:

By applying the phase rotation factor *<sup>v</sup> <sup>j</sup>*

where *<sup>T</sup>*

**3.2 Conventional PTS (C-PTS)** 

*V*

*v v 1*

*X X*

the lowest PAPR.

and Lee, 2005).

Fig. 2. Block diagram of the C-PTS scheme with Digital predistortion and power amplifier in series

This process is performed by choosing the optimization parameter *b* which satisfies the following condition:

$$\tilde{b} = \arg\min \{ \max\_{0 \le k \le N\mathcal{F}-1} \left| \sum\_{v=1}^{V} b\_v x\_v \right| \}\tag{10}$$

where *V* is the number of subblocks partitioning and *F* is the oversampling factor. After obtaining the optimum *b* , the signal is transmitted.

For finding the optimum *b* , we should perform exhaustive search for (*V-1*) phase factors since one phase factor can remain fixed, *b1=*1. Hence to find the optimum phase factor, *WV-1* iteration should be performed, where *W* is the number of allowed phase factors.

#### **3.3 Enhanced PTS (EPTS)**

In order to decrease the complexity of C-PTS, a new phase sequence is generated. The block diagram of the enhanced partial transmit sequence (EPTS) scheme is shown in Fig. 3.

This new phase sequence is based on the generation of *N* random values of {1 -1 j –j} if the allowed phase factors is *W*=4. The phase sequence matrix can be given by:

Peak-to-Average Power Ratio Reduction in

Fig. 3. The block diagram of enhanced PTS

Therefore the algorithm can be expressed as follows:

the PAPR reduction is less.

Orthogonal Frequency Division Multiplexing Systems 223

where (13) and (14) are the interleaved and adjacent phase sequences matrix, respectively. As an example take the case of *N*=256, and the number of allowed phase factor and subblock partitioning are *W*=4 and *V*=4 respectively. With C-PTS there are *WM-1*=64 possible iterations, whereas for the proposed method, in the case of *D=2*, the phase sequence is a matrix of [128x256] elements according to (11). In this case 64 iterations are required for finding the optimum phase sequence, because each two rows of the matrix in (11) multiply

The reduction of subblocks to 2 is because it gives almost the same PAPR reduction as C-PTS with *V*=4. It should be noted that if *D*=1 then the complexity increases while if D>2 then

Step 1: Generate the input data stream and map it to the M-QAM modulation. Step 2: Construct a matrix of random phase sequence with dimension of *[PxN]*.

Step 4: Find the optimum phase sequence after *P* iterations to minimize the PAPR.

Step 3: Point-wise multiply signal *xv* with the new phase sequence.

point-wise with the time domain input signal *xv* with length [2x256].

$$
\hat{b} = \begin{bmatrix} b\_{1,1} & & \dots & & b\_{1,N} \\ \vdots & & \vdots & & \vdots \\ b\_{V,1} & & \dots & & b\_{2,N} \\ b\_{V+1,1} & & \dots & & b\_{V+1,N} \\ \vdots & & \vdots & & \vdots \\ b\_{P,1} & \dots & & b\_{P,N} \end{bmatrix}\_{\mid P \sim \mathcal{N}} \tag{11}
$$

where *P* is the number of iterations that should be set in accordance with the number of iterations of the C-PTS and *N* is the number of samples (IFFT length) and *V* is the number of subblock partitioning. The value of P is given as follows:

$$P = D \mathcal{W}^{V-1} \qquad , \quad D = 1, 2, \dots, D\_N \tag{12}$$

where *D* is the coefficient that can be specified based on the PAPR reduction and complexity requirement and *DN* is specified by the user. The value of *P* explicitly depends on the number of subblocks *V*, if the number of allowed phase factor remains constant.

There is a tradeoff for choosing the value of *D*. higher *D* leads to higher PAPR reduction but at the expense of higher complexity; while lower *D* results in smaller PAPR reduction but with less complexity. For example if *W=*2 and *V=*4, then in C-PTS there are 8 iterations and hence *P=*8*D*. If *D=*2, then *P=*16 and both methods have the same number of iterations. But when *D=*1, then number of iterations to find the optimum phase factor will be reduced to 4 and this will result in complexity reduction. The main advantage of this method over C-PTS is the reduction of complexity while at the same time maintaining the same PAPR performance. In the case of C-PTS, each row of the matrix *ˆ b* contains same phase sequence while each column is periodical with period *V*, whereas in the proposed method each element of matrix *ˆ b* has different random values. The other formats that matrix in (11) can be expressed are as follows:

$$\hat{\boldsymbol{b}} = \begin{bmatrix} \hat{\boldsymbol{b}}\_{1,1}, \dots, \hat{\boldsymbol{b}}\_{1,N/P} \boldsymbol{P}\_{1}, \dots, \dots, \hat{\boldsymbol{b}}\_{1,T}, \dots, \hat{\boldsymbol{b}}\_{1,N/P} \\ \vdots & \vdots & \vdots \\ \boldsymbol{b}\_{V,1}, \dots, \hat{\boldsymbol{b}}\_{V,N/P} \boldsymbol{P}\_{1}, \dots, \dots \hat{\boldsymbol{b}}\_{V,1}, \dots, \hat{\boldsymbol{b}}\_{V,N/P} \\ \boldsymbol{b}\_{V+1,1}, \dots, \hat{\boldsymbol{b}}\_{V+1,N/P}, \dots, \hat{\boldsymbol{b}}\_{V+1,1}, \dots, \hat{\boldsymbol{b}}\_{V+1,N/P} \\ \vdots & \vdots & \vdots \\ \boldsymbol{b}\_{P,1}, \dots, \hat{\boldsymbol{b}}\_{P,N/P} & \dots, \dots \hat{\boldsymbol{b}}\_{P,1}, \dots, \hat{\boldsymbol{b}}\_{P,N/P} \end{bmatrix}\_{\mid P \gg 1} \tag{13}$$
 
$$\hat{\boldsymbol{b}} = \begin{bmatrix} \frac{p}{\hat{\boldsymbol{b}}\_{1,1}, \dots, \hat{\boldsymbol{b}}\_{1,1}}, \frac{p}{\hat{\boldsymbol{b}}\_{1,2}, \dots, \hat{\boldsymbol{b}}\_{1,2}}, \dots, \overbrace{\boldsymbol{b}\_{1,N/P}, \dots, \hat{\boldsymbol{b}}\_{1,N/P}}\_{\mid N/P, \dots, \hat{\boldsymbol{b}}\_{1,N/P}} \\ \vdots & \vdots & \vdots \\ \boldsymbol{b}\_{V,1}, \dots, \boldsymbol{b}\_{V,1}, \dots, \hat{\boldsymbol{b}}\_{V,2}, \dots, \dots \boldsymbol{b}\_{V,N/P}, \dots, \hat{\boldsymbol{b}}\_{V,N/P} \end{bmatrix}\_{\mid P \gg 1} \tag{14}$$

*1,1 1,N*

*b ,..., b*

*V ,1 2,N V 1,1 V 1,N*

*b ,..., b <sup>ˆ</sup> <sup>b</sup> b ,..., b*

subblock partitioning. The value of P is given as follows:

PTS, each row of the matrix *ˆ*

*P,1 P,N*

*V 1 P DW , D 1,2,...,DN*

number of subblocks *V*, if the number of allowed phase factor remains constant.

with period *V*, whereas in the proposed method each element of matrix *ˆ*

values. The other formats that matrix in (11) can be expressed are as follows:

*P 1,1 1,N / P 1,1 1,N / P*

 

*b ,...,b ,..., b ,...,b*

*ˆ b ,...,b ,..., b ,...,b b b ,...,b ,..., b ,...,b*

*V ,1 V ,N / P V ,1 V ,N / P V 1,1 V 1,N / P V 1,1 V 1,N / P*

*P ,1 P,N / P P ,1 P ,N / P*

*PP P 1,1 1,1 1,2 1,2 1,N / P 1,N / P*

 

*b ,...,b , b ,...,b ,..., b ,...,b*

*V ,1 V ,1 V ,2 V ,2 V ,N / P V ,N / P V 1,1 V 1,1 V 1,N / P V 1,N / P*

*b ,...,b ,..., b ,...,b*

*P,1 P,1 P,N / P*

*b ,...,b ,..., b ,...,b*

*ˆ b ,...,b ,b ,...,b ,...,b ,...,b b b ,...,b ,..., b ,...,b*

where *D* is the coefficient that can be specified based on the PAPR reduction and complexity requirement and *DN* is specified by the user. The value of *P* explicitly depends on the

There is a tradeoff for choosing the value of *D*. higher *D* leads to higher PAPR reduction but at the expense of higher complexity; while lower *D* results in smaller PAPR reduction but with less complexity. For example if *W=*2 and *V=*4, then in C-PTS there are 8 iterations and hence *P=*8*D*. If *D=*2, then *P=*16 and both methods have the same number of iterations. But when *D=*1, then number of iterations to find the optimum phase factor will be reduced to 4 and this will result in complexity reduction. The main advantage of this method over C-PTS is the reduction of complexity while at the same time maintaining the same PAPR performance. In the case of C-

where *P* is the number of iterations that should be set in accordance with the number of iterations of the C-PTS and *N* is the number of samples (IFFT length) and *V* is the number of

*b ,..., b*

*[P N]*

(12)

*b* contains same phase sequence while each column is periodical

*[ P*

*[P N] P,N / P*

*N ]*

(11)

*b* has different random

(14)

(13)

where (13) and (14) are the interleaved and adjacent phase sequences matrix, respectively.

As an example take the case of *N*=256, and the number of allowed phase factor and subblock partitioning are *W*=4 and *V*=4 respectively. With C-PTS there are *WM-1*=64 possible iterations, whereas for the proposed method, in the case of *D=2*, the phase sequence is a matrix of [128x256] elements according to (11). In this case 64 iterations are required for finding the optimum phase sequence, because each two rows of the matrix in (11) multiply point-wise with the time domain input signal *xv* with length [2x256].

Fig. 3. The block diagram of enhanced PTS

The reduction of subblocks to 2 is because it gives almost the same PAPR reduction as C-PTS with *V*=4. It should be noted that if *D*=1 then the complexity increases while if D>2 then the PAPR reduction is less.

Therefore the algorithm can be expressed as follows:

Step 1: Generate the input data stream and map it to the M-QAM modulation.

Step 2: Construct a matrix of random phase sequence with dimension of *[PxN]*.

Step 3: Point-wise multiply signal *xv* with the new phase sequence.

Step 4: Find the optimum phase sequence after *P* iterations to minimize the PAPR.

Peak-to-Average Power Ratio Reduction in

Orthogonal Frequency Division Multiplexing Systems 225

Fig. 5. CCDF comparison of PAPR of the proposed EPTS and C-PTS

The DSI method reduces PAPR by increasing the average power of the signal. Here, after converting the input data stream into parallel through the serial to parallel converter a, dummy sequence is inserted in the input signal. Therefore, the average value in Equation (2) is increased and the PAPR is subsequently reduced (Ryu, et al. 2004). IEEE 802.16d standard, specifies that the data frame of OFDM signal is allocated with 256 subcarriers which is composed of 192 data subcarriers, 1 zero DC subcarrier, 8 pilot subcarriers, and 55 guard subcarriers. Therefore, the dummy sequence can be inserted within the slot of 55 guard subcarriers without degradation of user data. However, if added dummies are more than 55, the length of the data and the bandwidth required, will be increased. This will degrade the

<sup>=</sup>*<sup>K</sup> TE ×100%*

where *K* is the number of the subcarriers and *L* is the number of dummy sequence. In this chapter we apply a different DSI method from the one in (Ryu, et al. 2004), where the *TE* is

*K+L* (15)

**3.4 Dummy Sequence Insertion (DSI)** 

Transmission Efficiency (*TE)* which is defined as:

always 100%.

#### **3.3.1 Numerical analysis**

In order to evaluate and compare the performance of the PAPR methods with C-PTS, simulations have been performed. In all the simulations, we employed QPSK modulation with IFFT length of *N=*512, and oversampling factor *F=*4. To obtain the complementary cumulative distribution function (CCDF), 40000 random OFDM symbols are generated**.** 

Fig. 4 shows the CCDF of three different types of phase sequences interleaved, adjacent and random for D=2. From this figure, PAPR reduction with random phase sequence outperforms the other types and hence this type of phase sequence is applied in the following simulations.

Fig. 4. CCDF of PAPR of the proposed method for different phase sequence when D=2

Fig. 5 shows the CCDF comparison of the PAPR of the C-PTS and EPTS for V=2 and 4. It is clear that the proposed EPTS shows better PAPR performance compared to C-PTS where almost 0.3 dB reduction is achieved with EPTS.

In order to evaluate and compare the performance of the PAPR methods with C-PTS, simulations have been performed. In all the simulations, we employed QPSK modulation with IFFT length of *N=*512, and oversampling factor *F=*4. To obtain the complementary cumulative distribution function (CCDF), 40000 random OFDM symbols are generated**.** 

Fig. 4 shows the CCDF of three different types of phase sequences interleaved, adjacent and random for D=2. From this figure, PAPR reduction with random phase sequence outperforms the other types and hence this type of phase sequence is applied in the

Fig. 4. CCDF of PAPR of the proposed method for different phase sequence when D=2

almost 0.3 dB reduction is achieved with EPTS.

Fig. 5 shows the CCDF comparison of the PAPR of the C-PTS and EPTS for V=2 and 4. It is clear that the proposed EPTS shows better PAPR performance compared to C-PTS where

**3.3.1 Numerical analysis** 

following simulations.

Fig. 5. CCDF comparison of PAPR of the proposed EPTS and C-PTS

#### **3.4 Dummy Sequence Insertion (DSI)**

The DSI method reduces PAPR by increasing the average power of the signal. Here, after converting the input data stream into parallel through the serial to parallel converter a, dummy sequence is inserted in the input signal. Therefore, the average value in Equation (2) is increased and the PAPR is subsequently reduced (Ryu, et al. 2004). IEEE 802.16d standard, specifies that the data frame of OFDM signal is allocated with 256 subcarriers which is composed of 192 data subcarriers, 1 zero DC subcarrier, 8 pilot subcarriers, and 55 guard subcarriers. Therefore, the dummy sequence can be inserted within the slot of 55 guard subcarriers without degradation of user data. However, if added dummies are more than 55, the length of the data and the bandwidth required, will be increased. This will degrade the Transmission Efficiency (*TE)* which is defined as:

$$\text{TE} = \frac{K}{K+L} \times 100\% \tag{15}$$

where *K* is the number of the subcarriers and *L* is the number of dummy sequence. In this chapter we apply a different DSI method from the one in (Ryu, et al. 2004), where the *TE* is always 100%.

Peak-to-Average Power Ratio Reduction in

where *<sup>T</sup>*

After finding the optimum *b*

is less than the *PAPRth* .

such that

condition:

Orthogonal Frequency Division Multiplexing Systems 227

As for the DSI-PTS method, consider *L* as the number of dummy sequence which later will be shown to be *L 55* and *N* is the IFFT length which is 256 in the case of fixed WiMAX that includes 192 data carriers, 8 pilots and 55 zero padding and 1 dc subcarrier. Here

From the block diagram in Fig. 6, *X* is the input signal stream with length *N* after which the dummy sequence is added. The dummy sequence can be replaced with zeros in data sample. This makes the IFFT length remain unchanged and decoding of the samples in

*S = [S ,S ,...,S ] v 12 V*

*S S*

and then these subblocks are combined to minimize the PAPR in time domain. In time domain the signal *vs* is oversampled *F* times which is obtained by taking an IFFT of length *FN* on signal *Xv* concatenated with *(F 1)N* zeros. After partitioning the signal and

*V*

the lowest PAPR. Notice that here *NKL* which means that there is no change in the length of the input signal after the addition of dummy sequence. The subblock partition type here is based on interleaving which is the best choice for PTS OFDM in terms of computational complexity reduction as compared to adjacent and pseudo-random method,

*b argmin( max b s )*

Then the PAPR of*s(b)* is checked whether it lies in the range of the PAPR threshold ( *PAPRth* ). After this additional task, the signal is transmitted otherwise it is returned to the DSI block to generate the dummy sequence again. This process will continue until the PAPR

Fig. 7 shows the CCDF curves of conventional PTS and DSI-PTS techniques. We assume here that the number of dummy sequence insertion ( *L* ) is 55 which bears no significant

*v v v 1 s(b) b s* 

*0 1 NF 1 s (b) [s (b),s (b),...s (b)]* . The objective is to find the optimum signal*s(b)* with

*V v v,k 0 k NF 1 v 1*

(18)

then the optimum signal *s(b)* is transmitted to the next block.

*<sup>v</sup> b e ,v 1,2,...,V* are used to

with the following

(17)

*V v v 1*

 

complementary sequence is applied for the DSI (Ryu, et al. 2004).

performing the IFFT for each part, then the phase factors *<sup>v</sup> <sup>j</sup>*

however it gives the least PAPR reduction among them.

optimize the *Sv* . In time domain the OFDM signal can be expressed as:

Then, the process is continued by choosing the optimization parameter *b*

receiver becomes simpler. Then the signal is partitioned into *V* disjoint blocks

#### **3.5 Dummy Sequence Insertion with Partial Transmit Sequence (DSI-PTS)**

The block diagram of this technique is shown in Fig. 6. A complex valued dummy signals are first generated and then added to the vector of data subcarriers. The new vector in frequency domain is then constructed from K-data and L-dummy subcarriers, respectively. L can be any number less than K. The new vector S is given by:

$$S = \left[ X\_{k'} W\_l \right] \tag{16}$$

where *X [X ,X ,...,X ],k 1,2,...,K k k ,0 k ,1 k ,N L 1* is the data subcarrier vector and *W [W ,W ,...,W ],l 1,2,...,L l l,0 l,1 l,L 1* is the dummy signals vector.

After generation of the optimum OFDM signal then the PAPR is checked with the acceptable threshold that was pre-defined before. If the PAPR value is less than the threshold then the OFDM signal will be transmitted otherwise the dummy sequence is generated again as depicted with the feedback in Fig. 6. This process is one iteration. The number of iterations can be increased to achieve the desired PAPR ( *PAPRth* ) reduction but the processing time will also increase likewise and causes the system performance to drop.

Fig. 6. Block diagram of DSI-PTS technique

As for the DSI-PTS method, consider *L* as the number of dummy sequence which later will be shown to be *L 55* and *N* is the IFFT length which is 256 in the case of fixed WiMAX that includes 192 data carriers, 8 pilots and 55 zero padding and 1 dc subcarrier. Here complementary sequence is applied for the DSI (Ryu, et al. 2004).

From the block diagram in Fig. 6, *X* is the input signal stream with length *N* after which the dummy sequence is added. The dummy sequence can be replaced with zeros in data sample. This makes the IFFT length remain unchanged and decoding of the samples in receiver becomes simpler. Then the signal is partitioned into *V* disjoint blocks

$$S\_v = \{ \mathbf{S}\_1, \mathbf{S}\_2, \dots, \mathbf{S}\_V \}$$

such that

226 Advanced Transmission Techniques in WiMAX

The block diagram of this technique is shown in Fig. 6. A complex valued dummy signals are first generated and then added to the vector of data subcarriers. The new vector in frequency domain is then constructed from K-data and L-dummy subcarriers, respectively.

where *X [X ,X ,...,X ],k 1,2,...,K k k ,0 k ,1 k ,N L 1* is the data subcarrier vector and

After generation of the optimum OFDM signal then the PAPR is checked with the acceptable threshold that was pre-defined before. If the PAPR value is less than the threshold then the OFDM signal will be transmitted otherwise the dummy sequence is generated again as depicted with the feedback in Fig. 6. This process is one iteration. The number of iterations can be increased to achieve the desired PAPR ( *PAPRth* ) reduction but the processing time will also increase likewise and causes the system performance to

*S X ,W k l* (16)

**3.5 Dummy Sequence Insertion with Partial Transmit Sequence (DSI-PTS)** 

L can be any number less than K. The new vector S is given by:

*W [W ,W ,...,W ],l 1,2,...,L l l,0 l,1 l,L 1* is the dummy signals vector.

Fig. 6. Block diagram of DSI-PTS technique

drop.

$$\sum\_{v=1}^{V} S\_v = S$$

and then these subblocks are combined to minimize the PAPR in time domain. In time domain the signal *vs* is oversampled *F* times which is obtained by taking an IFFT of length *FN* on signal *Xv* concatenated with *(F 1)N* zeros. After partitioning the signal and performing the IFFT for each part, then the phase factors *<sup>v</sup> <sup>j</sup> <sup>v</sup> b e ,v 1,2,...,V* are used to optimize the *Sv* . In time domain the OFDM signal can be expressed as:

$$s'(b) = \sum\_{v=1}^{V} b\_v s\_v \tag{17}$$

where *<sup>T</sup> 0 1 NF 1 s (b) [s (b),s (b),...s (b)]* . The objective is to find the optimum signal*s(b)* with the lowest PAPR. Notice that here *NKL* which means that there is no change in the length of the input signal after the addition of dummy sequence. The subblock partition type here is based on interleaving which is the best choice for PTS OFDM in terms of computational complexity reduction as compared to adjacent and pseudo-random method, however it gives the least PAPR reduction among them.

Then, the process is continued by choosing the optimization parameter *b* with the following condition:

$$\tilde{b} = \arg\min \{ \max\_{0 \le k \le N\Gamma - 1} \left| \sum\_{v=1}^{V} b\_v s\_{v,k} \right|\}\tag{18}$$

After finding the optimum *b* then the optimum signal *s(b)* is transmitted to the next block. Then the PAPR of*s(b)* is checked whether it lies in the range of the PAPR threshold ( *PAPRth* ). After this additional task, the signal is transmitted otherwise it is returned to the DSI block to generate the dummy sequence again. This process will continue until the PAPR is less than the *PAPRth* .

Fig. 7 shows the CCDF curves of conventional PTS and DSI-PTS techniques. We assume here that the number of dummy sequence insertion ( *L* ) is 55 which bears no significant

Peak-to-Average Power Ratio Reduction in

I=10.

L=55

Orthogonal Frequency Division Multiplexing Systems 229

Fig. 8. CCDF of PAPR of DSI-PTS technique for different length of dummy sequence when

Fig. 9. CCDF of PAPR of DSI-PTS technique for different number of iterations when the

effect on the transmission efficiency (*TE = 100%* ). These results are obtained after 10 iteration (I). It can be observed that the PAPR reduction of our proposed PTS scheme outperforms the conventional PTS scheme with an improvement by 2 and 1 dB respectively at *CCDF = 0.01%* , when *V 2,4* respectively. Even though this reduction seems minor the complexity according to Table 1 is reduced significantly.

Fig. 7. CCDF of the PAPR of conventional PTS and DSI-PTS technique (L=55, I=10).

Fig. 8 shows the result for different length of dummy sequence. As discussed earlier the maximum length of dummy sequence that can be applied is 55 and this figure shows that with this length the reduction obtained is slightly better than when is 30. It is observed that the reductions of PAPR at *CCDF = 0.01%* are 1 dB, 1.5 dB and 2 dB for dummy length of 5, 30 and 55 respectively.

Fig. 9 shows the effect of different iteration number on the PAPR performance. From this figure maximum PAPR reduction is achieved which is 7 dB at *CCDF = 0.01%* at 100 iterations with *L = 55* . But increasing the number of iterations will reduce the data rate.

effect on the transmission efficiency (*TE = 100%* ). These results are obtained after 10 iteration (I). It can be observed that the PAPR reduction of our proposed PTS scheme outperforms the conventional PTS scheme with an improvement by 2 and 1 dB respectively at *CCDF = 0.01%* , when *V 2,4* respectively. Even though this reduction seems minor the

Fig. 7. CCDF of the PAPR of conventional PTS and DSI-PTS technique (L=55, I=10).

30 and 55 respectively.

Fig. 8 shows the result for different length of dummy sequence. As discussed earlier the maximum length of dummy sequence that can be applied is 55 and this figure shows that with this length the reduction obtained is slightly better than when is 30. It is observed that the reductions of PAPR at *CCDF = 0.01%* are 1 dB, 1.5 dB and 2 dB for dummy length of 5,

Fig. 9 shows the effect of different iteration number on the PAPR performance. From this figure maximum PAPR reduction is achieved which is 7 dB at *CCDF = 0.01%* at 100 iterations with *L = 55* . But increasing the number of iterations will reduce the data rate.

complexity according to Table 1 is reduced significantly.

Fig. 8. CCDF of PAPR of DSI-PTS technique for different length of dummy sequence when I=10.

Fig. 9. CCDF of PAPR of DSI-PTS technique for different number of iterations when the L=55

Peak-to-Average Power Ratio Reduction in

which causes an increase in the IFFT length.

Fig. 11. Block diagram of the proposed DSI-EPTS scheme

*V 1 T 3VN / 2lo C PTS gN 2VW N*

where *P* is the number of iterations and *V* is the number of subblocks.

in section 3.5.

Zhou, 2007):

**3.6.1 Computational complexity** 

Whereas for the Enhanced PTS this value is:

Orthogonal Frequency Division Multiplexing Systems 231

shown by the feedback loop in Fig. 11. This process is one iteration. The number of iterations can be increased to achieve the desired PAPR ( *PAPRth* ) reduction but the processing time will also increase likewise and cause the system performance to drop. From the block diagram in Fig. 11, *X* is the input signal with length *N*. After that dummy sequence is added

The same procedure similar to the one discussed in section 3.5 for DSI-PTS scheme is performed here except the phase sequence is taken from the EPTS scheme discussed earlier

The total complexity of the C-PTS with oversampling factor *F=*1, is given by (Baxley and

In (Varahram, et al. 2010), the complexity is calculated only for IFFT section, but here we require the total complexity. Hence the total complexity for the DSI-PTS method is given by:

(20)

*T 3 / 4VN log N PVN EPTS* (21)

There is about 0.5 dB improvement in PAPR reduction when the number of iteration is 100 compared to 10 iteration for both cases of *V 2,4* as shown in Fig. 9.

Fig. 10. CCDF of PAPR of DSI-PTS technique compared to DSI when the number of iterations is 10 and *V*=2.

Fig. 10 demonstrates the PAPR reduction capacity in DSI and DSI-PTS techniques. It should also be highlighted on that. The DSI-PTS technique offers about 1.5 dB further reduction in PAPR compared to DSI when the number of dummy sequence *L = 55* and *V 2* .

#### **3.6 Dummy Sequence Insertion with Enhanced Partial Transmit Sequence (DSI-EPTS)**

The block diagram of this technique is shown in Fig. 11. Here as in DSI described previously, the complex valued dummy signals are first generated and then added to the vector of data subcarriers. The new vector in the frequency domain is then constructed from *K*-data and *L*-dummy subcarriers, respectively. *L* can be any number less than *K*. The new vector *U* is given by:

$$\mathcal{U} = \left\lfloor X\_k, \mathcal{W}\_l \right\rfloor \tag{19}$$

where *X [X ,X ,...,X ], k 1,2,...,K k k ,0 k ,1 k ,N L 1* is the data subcarrier vector and *W [W ,W ,...,W ], l 1,2,...,L l l,0 l,1 l,L 1* is the dummy signals vector.

After generation of the optimum OFDM signal, PAPR is checked with the acceptable threshold that has been predefined before. If the PAPR value is less than the threshold then the OFDM signal will be transmitted otherwise the dummy sequence is generated again as

There is about 0.5 dB improvement in PAPR reduction when the number of iteration is 100

Fig. 10. CCDF of PAPR of DSI-PTS technique compared to DSI when the number of

PAPR compared to DSI when the number of dummy sequence *L = 55* and *V 2* .

Fig. 10 demonstrates the PAPR reduction capacity in DSI and DSI-PTS techniques. It should also be highlighted on that. The DSI-PTS technique offers about 1.5 dB further reduction in

**3.6 Dummy Sequence Insertion with Enhanced Partial Transmit Sequence (DSI-EPTS)**  The block diagram of this technique is shown in Fig. 11. Here as in DSI described previously, the complex valued dummy signals are first generated and then added to the vector of data subcarriers. The new vector in the frequency domain is then constructed from *K*-data and *L*-dummy subcarriers, respectively. *L* can be any number less than *K*. The new

*U X ,W k l* (19)

where *X [X ,X ,...,X ], k 1,2,...,K k k ,0 k ,1 k ,N L 1* is the data subcarrier vector and

After generation of the optimum OFDM signal, PAPR is checked with the acceptable threshold that has been predefined before. If the PAPR value is less than the threshold then the OFDM signal will be transmitted otherwise the dummy sequence is generated again as

*W [W ,W ,...,W ], l 1,2,...,L l l,0 l,1 l,L 1* is the dummy signals vector.

iterations is 10 and *V*=2.

vector *U* is given by:

compared to 10 iteration for both cases of *V 2,4* as shown in Fig. 9.

shown by the feedback loop in Fig. 11. This process is one iteration. The number of iterations can be increased to achieve the desired PAPR ( *PAPRth* ) reduction but the processing time will also increase likewise and cause the system performance to drop. From the block diagram in Fig. 11, *X* is the input signal with length *N*. After that dummy sequence is added which causes an increase in the IFFT length.

Fig. 11. Block diagram of the proposed DSI-EPTS scheme

The same procedure similar to the one discussed in section 3.5 for DSI-PTS scheme is performed here except the phase sequence is taken from the EPTS scheme discussed earlier in section 3.5.

#### **3.6.1 Computational complexity**

The total complexity of the C-PTS with oversampling factor *F=*1, is given by (Baxley and Zhou, 2007):

$$T\_{\text{C-PTS}} = 3 \text{VN} \;/\; 2 \log \text{N} + 2 \text{V} \text{W}^{V-1} \text{N} \tag{20}$$

Whereas for the Enhanced PTS this value is:

$$T\_{EPTS} = 3 \nearrow 4 \text{VN} \log N + PVN \tag{21}$$

where *P* is the number of iterations and *V* is the number of subblocks.

In (Varahram, et al. 2010), the complexity is calculated only for IFFT section, but here we require the total complexity. Hence the total complexity for the DSI-PTS method is given by:

Peak-to-Average Power Ratio Reduction in

improves, as shown in the simulations.

of spectrum efficiency degradation.

**3.6.2 Side information** 

space is required.

follows:

**3.6.3 System performance** 

reduction and complexity according to equation (12).

The number of required side information bits in C-PTS is

power amplifiers and decreases the cost of the system.

as C-PTS for V=4 and V=8 respectively.

Orthogonal Frequency Division Multiplexing Systems 233

Table 1 presents the computational complexity of C-PTS and DSI-PTS, for *N=*512 and *W=*2. Table 2 presents the computational complexity of C-PTS and proposed DSI-EPTS, for the same value of *N* and *W,* while D is the coefficient that can be specified based on the PAPR

It is clear from Table 2, that CCRR is improved for both *V=4* and *V=8*. It should be noted that when *D* increases, the complexity reduction becomes less while PAPR performance

The other important factor in studying the PAPR reduction method is the side information which has to be transmitted to the receiver to extract the original signal. One method is that the side information can be transmitted in a separate channel but this comes at the expense

> *V 1 <sup>2</sup> log W*

where *W* is the number of allowed phase factors and the sign indicates the floor of y. In DSI-EPTS, the side information can be allocated in the dummy signals and therefore does not have impact on spectrum efficinecy and data rate loss; however, the only drawback of this method is that, because of the increase in the phase sequence matrix, higher memory

In C-PTS, even though an OFDM signal does not experience distortion the signal after power amplifier could exhibit distortions if PAPR is higher than the expected value. In this case the power amplifier should back off which degrades the efficiency of the system. In DSI-EPTS, the addition of dummy sequences causes the transmission efficiency to change as

where *K* is the length of subcarriers and *L* is the length of dummy sequences. In actual applications where the cost of the system is the main issue, the other block also have to be considered, the digital predistortion (DPD) (Varahram and Atlasbaf, 2005), (Varahram, et al. 2005). By applying DPD technique, it is possible to increase the linearity of the power amplifier and as a result, higher peak signals can be transmitted by the power amplifier and the performance of the PAPR can be improved. This also increases the efficiency of the

Fig. 12 shows the CCDF comparison of PAPR of DSI-EPTS with C-PTS. It is clear that by applying the DSI-EPTS when D=2, the PAPR performance is more superior over that of C-PTS for both V=4 and V=8 respectively. But PAPR reduction when D=1 is almost the same

*<sup>K</sup> TE 100[%] K L* (25)

$$T\_{\rm DSI-PTS} = 3 \, / \, 4 \text{VN} \log \text{N} + 2 \text{VNN} \text{V}^{V-1} + \text{QL} \tag{22}$$

The total complexity of DSI-EPTS is given by;

$$T\_{DS1-EPS} = 3 \ne 4 \text{VN} \log \text{N} + \text{PVN} + \text{QL} \tag{23}$$

where *Q* is the number of iterations for the DSI loop.

It can be observed that (22) and (23) consist of two parts; the first part is actually the complexity of the IFFT itself and the second part is the complexity of the searching algorithm. Most of the papers did not consider the second part which causes wrong calculation of the complexity. It should be noted that the number of IFFT in (24) and (25) is halved which basically is concluded from the simulation results. From the simulation results given in the following section the PAPR performance of the proposed method when the number of IFFT is half of the C-PTS, is almost the same. This is shown for different number of subblocks which proves that in the DSI-EPTS the number of IFFT is halved compared to the C-PTS but gives the same PAPR performance.


Table 1. Computational Complexity of the DSI-PTS and the conventional PTS when *N*=512 and *W*=2, *Q*=3, *L*=56


Table 2. Computational Complexity of the DSI-EPTS and the conventional PTS when *N*=512 and *W*=2, *Q*=3, *L*=56

The computational complexity reduction ratio (CCRR) of the proposed technique over the C-PTS is defined as (Baxley and Zhou, 2007):

$$\text{CCRR} = (1 - \frac{\text{Complexity of the DSI} - \text{EPTS}}{\text{Complexity of the C} - \text{PTS}}) \times 100\% \tag{24}$$

Table 1 presents the computational complexity of C-PTS and DSI-PTS, for *N=*512 and *W=*2.

Table 2 presents the computational complexity of C-PTS and proposed DSI-EPTS, for the same value of *N* and *W,* while D is the coefficient that can be specified based on the PAPR reduction and complexity according to equation (12).

It is clear from Table 2, that CCRR is improved for both *V=4* and *V=8*. It should be noted that when *D* increases, the complexity reduction becomes less while PAPR performance improves, as shown in the simulations.

#### **3.6.2 Side information**

232 Advanced Transmission Techniques in WiMAX

It can be observed that (22) and (23) consist of two parts; the first part is actually the complexity of the IFFT itself and the second part is the complexity of the searching algorithm. Most of the papers did not consider the second part which causes wrong calculation of the complexity. It should be noted that the number of IFFT in (24) and (25) is halved which basically is concluded from the simulation results. From the simulation results given in the following section the PAPR performance of the proposed method when the number of IFFT is half of the C-PTS, is almost the same. This is shown for different number of subblocks which proves that in the DSI-EPTS the number of IFFT is halved compared to

Table 1. Computational Complexity of the DSI-PTS and the conventional PTS when *N*=512

Table 2. Computational Complexity of the DSI-EPTS and the conventional PTS when *N*=512

The computational complexity reduction ratio (CCRR) of the proposed technique over the

*Complexity of theDSI EPTS CCRR (1 ) Complexity of theC PTS*

The total complexity of DSI-EPTS is given by;

where *Q* is the number of iterations for the DSI loop.

the C-PTS but gives the same PAPR performance.

No. of

C-PTS is defined as (Baxley and Zhou, 2007):

Subblocks C-PTS

Total Complexity

and *W*=2, *Q*=3, *L*=56

Total Complexity

and *W*=2, *Q*=3, *L*=56

No. of

*V 1 T 3 / 4VN lo DSI PTS gN 2VNW QL* (22)

*T 3 / 4VN log N PVN QL DSI EPTS* (23)

Subblocks C-PTS DSI-PTS CCRR

V=4 60416 46760 22.6%

V=8 1103872 1076392 2.4%

V=4 60416 30376 46760 49.7 22.6

V=8 1103872 552104 1076392 49.6 2.7

*100%* (24)

DSI-EPTS CCRR (%)

D=1 D=2 D=1 D=2

The other important factor in studying the PAPR reduction method is the side information which has to be transmitted to the receiver to extract the original signal. One method is that the side information can be transmitted in a separate channel but this comes at the expense of spectrum efficiency degradation.

The number of required side information bits in C-PTS is

$$
\left\lfloor \log\_2 \boldsymbol{W}^{V-1} \right\rfloor
$$

where *W* is the number of allowed phase factors and the sign indicates the floor of y. In DSI-EPTS, the side information can be allocated in the dummy signals and therefore does not have impact on spectrum efficinecy and data rate loss; however, the only drawback of this method is that, because of the increase in the phase sequence matrix, higher memory space is required.

#### **3.6.3 System performance**

In C-PTS, even though an OFDM signal does not experience distortion the signal after power amplifier could exhibit distortions if PAPR is higher than the expected value. In this case the power amplifier should back off which degrades the efficiency of the system. In DSI-EPTS, the addition of dummy sequences causes the transmission efficiency to change as follows:

$$\text{TE} = \frac{\text{K}}{\text{K} + \text{L}} \times 100 \text{[\%]} \tag{25}$$

where *K* is the length of subcarriers and *L* is the length of dummy sequences. In actual applications where the cost of the system is the main issue, the other block also have to be considered, the digital predistortion (DPD) (Varahram and Atlasbaf, 2005), (Varahram, et al. 2005). By applying DPD technique, it is possible to increase the linearity of the power amplifier and as a result, higher peak signals can be transmitted by the power amplifier and the performance of the PAPR can be improved. This also increases the efficiency of the power amplifiers and decreases the cost of the system.

Fig. 12 shows the CCDF comparison of PAPR of DSI-EPTS with C-PTS. It is clear that by applying the DSI-EPTS when D=2, the PAPR performance is more superior over that of C-PTS for both V=4 and V=8 respectively. But PAPR reduction when D=1 is almost the same as C-PTS for V=4 and V=8 respectively.

Peak-to-Average Power Ratio Reduction in

performance and complexity reduction.

outperforms DSI-PTS for both V=2 and V=4 respectively.

system shows improvement at the cost of BER.

AWGN channels.

Orthogonal Frequency Division Multiplexing Systems 235

The highest PAPR reduction is achieved when D=2 and V=4. From table 2, the complexity reduction is minimum when D=2. There is always a trade off between PAPR reduction

Fig. 13 shows the CCDF comparison of PAPR of the DSI-EPTS and DSI-PTS when L=56. The results are shown for V=2 and V=4. The CCDF results show that PAPR of the DSI-EPTS

Fig. 14 shows a comparison of Bit Error Rate (BER) performance of the conventional PTS and the proposed EPTS and DSI-EPTS method in Additive White Gaussian Noise (AWGN) channels. The length of dummy sequence and iterations is L=56. From this figure, we can see that the BER is slightly increased with DSI-EPTS method compared to conventional PTS, but PAPR is much improved according to the result of Fig. 13. The performance of the

Fig. 14**.** Comparison of BER performance of the conventional PTS and DSI-PTS technique in

Fig. 12. CCDF comparison of PAPR of the DSI-EPTS and C-PTS

Fig. 13. CCDF comparison of PAPR of the DSI-EPTS and DSI-PTS when L=56

Fig. 12. CCDF comparison of PAPR of the DSI-EPTS and C-PTS

Fig. 13. CCDF comparison of PAPR of the DSI-EPTS and DSI-PTS when L=56

The highest PAPR reduction is achieved when D=2 and V=4. From table 2, the complexity reduction is minimum when D=2. There is always a trade off between PAPR reduction performance and complexity reduction.

Fig. 13 shows the CCDF comparison of PAPR of the DSI-EPTS and DSI-PTS when L=56. The results are shown for V=2 and V=4. The CCDF results show that PAPR of the DSI-EPTS outperforms DSI-PTS for both V=2 and V=4 respectively.

Fig. 14 shows a comparison of Bit Error Rate (BER) performance of the conventional PTS and the proposed EPTS and DSI-EPTS method in Additive White Gaussian Noise (AWGN) channels. The length of dummy sequence and iterations is L=56. From this figure, we can see that the BER is slightly increased with DSI-EPTS method compared to conventional PTS, but PAPR is much improved according to the result of Fig. 13. The performance of the system shows improvement at the cost of BER.

Fig. 14**.** Comparison of BER performance of the conventional PTS and DSI-PTS technique in AWGN channels.

Peak-to-Average Power Ratio Reduction in

1727, Oct. 2010.

89-94, Feb. 2004

no. 3, 2009.

2005.

pp. 25, 2009.

*ECCTD*, pp. 488 – 491, 2007.

Academic Pub, 2000.

Orthogonal Frequency Division Multiplexing Systems 237

Mohammady S., Varahram P., Sidek R. M., Hamidon M. N. and Sulaiman N., "Efficiency

Muller S. H. and Huber J. B., "A novel peak power reduction scheme for OFDM," in *Personal,* 

Nikookar H. and Lidsheim K. S., "Random phase updating algorithm for OFDM transmission with low PAPR," *IEEE Trans. Broadcast.,* vol. 48, pp. 123-128, 2002. Qian H., "Power Efficiency Improvements for Wireless Transmissions," *Power Efficiency* 

Ryu H. G., Lee J.E. and Park J.S., "Dummy sequence insertion (DSI) for PAPR reduction in

Tellado J., *Multicarrier Modulation with Low PAR: Applications to DSL and Wireless.* Kluwer

Tan C. E. and Wassell I. J., "Data bearing peak reduction carriers for OFDM systems,"

Van Nee R. and De Wild A., "Reducing the peak-to-average power ratio of OFDM," in *48th* 

Varahram P., Mohammady S., Hamidon M. N., Sidek R. M. and Khatun S., "Digital

Varahram P., Mohammady S., Hamidon M. N., Sidek R. M. and Khatun S., "Power

Varahram P., Atlasbaf Z., "Adaptive digital predistortion for high power amplifiers with

Varahram P., Azzo W. A., Ali B. M., "A Low Complexity Partial Transmit Sequence Scheme

Varahram P., Atlasbaf Z., Heydarian N., "Adaptive digital predistortion for power

Vijayarangan V. and Sukanesh D., "AN OVERVIEW OF TECHNIQUES FOR REDUCING

*International Conference on,* vol. 2; 2, pp. 854-858 vol.2, 2003.

*IEEE Vehicular Technology Conference, 1998. VTC 98,* 1998.

*Consumer Electron*, vol. 56, no. 4, pp. 2416-2420, Nov. 2010.

*Electromagnetics, APACE Proceedings,* (1607810):215-21, 2005.

*the 8th IEEE International Symposium on,* 1997.

*Improvements for Wireless Transmissions,* 2005.

improvement in microwave power amplifiers by using Complex Gain Predistortion technique," *IEICE Electronics Express (ELEX),* vol. 7, no. 23, pp.1721-

*Indoor and Mobile Radio Communications, 1997.'Waves of the Year 2000'. PIMRC'97.,* 

the OFDM communication system," *IEEE Trans. Consumer Electr.*, vol. 50, no. 1, pp.

*Information, Communications and Signal Processing, 2003 and the Fourth Pacific Rim Conference on Multimedia. Proceedings of the 2003 Joint Conference of the Fourth* 

Predistortion Technique for Compensating Memory Effects of Power Amplifiers in Wideband Applications", *Journal of Electrical Engineering*, vol. 60,

amplifiers linearization based on digital predistortion with memory effects used in CDMA applicationsm," *18th European Conference on Circuit Theory and Design* 

memory effects," *Asia-Pacific Microwave Conference Proceedings, APMC.* 3(1606627),

by Use of Dummy Signals for PAPR Reduction in OFDM Systems," *IEEE Trans.* 

amplifiers used in CDMA applications," *Asia-Pacific Conference on Applied* 

PEAK TO AVERAGE POWER RATIO AND ITS SELECTION CRITERIA FOR ORTHOGONAL FREQUENCY DIVISION MULTIPLEXING RADIO SYSTEMS,"

#### **4. Conclusion**

In this chapter we have studied and discussed several PAPR redcution techniques. Their advantages and disadvantages have been analyzed and by performing the simulation results, the PAPR performance of those techniques have been compared. Also the complexity of each technique has been computed and finally compared. These PAPR techniques is ideal for the latest wireless communications systems such as WiMAX and long term evolution (LTE).

#### **5. Acknowledgment**

This work was supported by Universiti Putra Malaysia under the Research University Grant Scheme (No. 0501090724RU).

#### **6. References**


In this chapter we have studied and discussed several PAPR redcution techniques. Their advantages and disadvantages have been analyzed and by performing the simulation results, the PAPR performance of those techniques have been compared. Also the complexity of each technique has been computed and finally compared. These PAPR techniques is ideal for the latest wireless communications systems such as WiMAX and long

This work was supported by Universiti Putra Malaysia under the Research University Grant

Al-Azoo W. F., Ali B. M, Khatun S., Bilfagih S. M, and Noordin N. K., "Addition of

Bauml v, Fischer R. F. H. and Huber J. B., "Reducing the peak-to-average power ratio of

Baxley R. J. and Zhou T., "Comparing slected mapping and partial transmit sequence for

Cimini L. J. and Sollenberger N. R., "Peak-to-average power ratio reduction of an OFDM

Han S. H. and Lee J. H., "An overview of peak-to-average power ratio reduction techniques

Hao M. and Liaw C., "A companding technique for PAPR reduction of OFDM systems,"

Foomooljareon P. and Fernando W. A. C., "Input sequence envelope scaling in PAPR

Jayalath v and Tellambura C.. (2000, Reducing the peak-to-average power ratio of

Krongold B. S. and Jones D. L., "PAR reduction in OFDM via active constellation extension,"

Kwon U., Kim D. and Im G., "Amplitude clipping and iterative reconstruction of MIMO-

*IEICE Trans. Commun.,* vol. E91-B, pp. 935-938, 2008.

interleaving. *Electron. Lett. 36(13),* pp. 1161-1163.

systems," *IEEE ICCCE'08 International conference*, PP. 1344-1347, 2008. Bauml R. W., Fischer R. F. H. and Huber J. B., "Reducing the peak-to-average power ratio of

Gaussian random signals for peak to average power ration reduction in OFDM

multicarrier modulationby selected mapping," *Electron. Lett.,* vol. 32, pp. 2056-2057,

multicarrier modulationby selected mapping," *Electron. Lett.,* vol. 32, pp. 2056-2057,

PAPR reduction", *IEEE Trans. Broadcast*., vol. 53, no. 4, pp.797-803, December

signal using partial transmit sequences," *IEEE Commun. Lett.*, vol. 4, no. 3, pp. 86–

for multicarrier transmission," *IEEE Wireless Commun.*, vol. 12, no. 2, pp. 56-65, Apr.

orthogonal frequencydivision multiplexing signal through bit or symbol

OFDM signals with optimum equalization," *IEEE Transactions on Wireless* 

**4. Conclusion** 

term evolution (LTE).

**5. Acknowledgment** 

1996.

1996.

2007.

2005.

88, March. 2000.

reduction of OFDM," in 2002.

*IEEE Trans. Broadcast.,* vol. 49, 2003.

*Communications,* vol. 8, pp. 268-277, 2009.

**6. References** 

Scheme (No. 0501090724RU).


**12** 

*China* 

**Design and Implementation of** 

*Beijing University of Posts and Telecommunications,*

Design and implementation of a wireless communication system protocol stack on the hardware platform are challenging tasks, which are seldom mentioned in published results currently. In fact, the work aims at achieving the predetermined function and performance based on the specific hardware resource, which involves how to design the software architecture according to hardware, how to choice the suitable algorithms and program them optimally on programmable DSP or embedded processor, etc. Typically, there are two ultimate application modes for a communication protocol stack: dedicated ASIC or programmable DSP. However, before the protocol stack is formed into dedicated ASIC, the task of protocol or algorithms implementation and testing on the programmable DSP should be finished primarily. Therefore, the chapter will focus on the topic of development of the

The first section, we will discuss the overall software and hardware architecture of a WiMAX TDD baseband system, and what are the most import considerations in this design

The second section, we address on the topic of developing the WiMAX PHY protocol on the

The third section, it is about the design and implementation of WiMAX MAC protocol on

Wireless systems designers need to meet a number of critical requirements including processing speed, flexibility, and time-to-market, all of which ultimately drive the hardware

WiMAX PHY/MAC protocol on a multi-core DSP platform.

programmable multi-core DSP platform.

**2.1 Baseband system 2.1.1 Requirements** 

platform choice.

**2. Hardware platform for WiMAX system** 

The contents of this chapter will be organized into three sections:

the embedded processor, on which embedded Linux OS is running.

**1. Introduction** 

phase.

 **WiMAX Baseband System** 

Zhuo Sun, Xu Zhu, Rui Chen, Zhuoyi Chen and Mingli Peng

Wilkison T. A. and Jones A. E., "Minimazation of the peak to mean envelope power ratio of multicarrier transmission schemes by block coding," in *IEEE Vehicular Technology Conference,* 1995.

### **Design and Implementation of WiMAX Baseband System**

Zhuo Sun, Xu Zhu, Rui Chen, Zhuoyi Chen and Mingli Peng *Beijing University of Posts and Telecommunications, China* 

#### **1. Introduction**

238 Advanced Transmission Techniques in WiMAX

Wilkison T. A. and Jones A. E., "Minimazation of the peak to mean envelope power ratio of

*Conference,* 1995.

multicarrier transmission schemes by block coding," in *IEEE Vehicular Technology* 

Design and implementation of a wireless communication system protocol stack on the hardware platform are challenging tasks, which are seldom mentioned in published results currently. In fact, the work aims at achieving the predetermined function and performance based on the specific hardware resource, which involves how to design the software architecture according to hardware, how to choice the suitable algorithms and program them optimally on programmable DSP or embedded processor, etc. Typically, there are two ultimate application modes for a communication protocol stack: dedicated ASIC or programmable DSP. However, before the protocol stack is formed into dedicated ASIC, the task of protocol or algorithms implementation and testing on the programmable DSP should be finished primarily. Therefore, the chapter will focus on the topic of development of the WiMAX PHY/MAC protocol on a multi-core DSP platform.

The contents of this chapter will be organized into three sections:

The first section, we will discuss the overall software and hardware architecture of a WiMAX TDD baseband system, and what are the most import considerations in this design phase.

The second section, we address on the topic of developing the WiMAX PHY protocol on the programmable multi-core DSP platform.

The third section, it is about the design and implementation of WiMAX MAC protocol on the embedded processor, on which embedded Linux OS is running.

#### **2. Hardware platform for WiMAX system**

#### **2.1 Baseband system**

#### **2.1.1 Requirements**

Wireless systems designers need to meet a number of critical requirements including processing speed, flexibility, and time-to-market, all of which ultimately drive the hardware platform choice.

Design and Implementation of WiMAX Baseband System 241

Fig. 1. The general hardware architecture of mobile communications system

The process in IF contains NCO (numerical controlled oscillator), CIC filter, half-band filter and FIR filter. All of the modules mentioned above have such a simple structure that they can be easily implemented in an FPGA. Because of the high data-rate of IF digital data, it is impossible to implemented in DSP. For example, first decimation filter in a digital wireless receiver, typically, is a CIC filter, operating at a sample rate of 50-100MHz. At these rates any DSP processor would find it extremely difficult to do anything. However, the CIC has an extremely simple structure, and implementing it in an FPGA would be easy. A sample rate of 100MHz should be achievable, and even the smallest FPGA will have a lot of resource available for further processing. In addition, the latest mobile communications system has employed MIMO technology, which means there will be two or more antennas. As the fact of introducing MIMO, the parallel sampling data will be processed simultaneously. Therefore, it is general to choose FPGA or ASIC as the processor in IF instead of DSP. For example, AD6654 is an IF to baseband receiver, with programmable decimating FIR filters, interpolating half-band filters and CIC filters built-in (Fathi, 2004).

Baseband is usually divided into two parts. One of them is digital signal processing, which is used for implementing PHY protocol. The other part is microprocessor used for implementing MAC and lower protocol (Goldfarb et al., 2000). The microprocessor is generally selected within ARM or Power PC processor. However, the choice of baseband digital signal processing solution is various. Generally there are four choices which are ASIC, DSP, FPGA and DSP+FPGA (Jusslia et al., 2001). Application-specific integrated circuit (ASIC), is an integrated circuit (IC) customized for a particular use, rather than intended for general-purpose use. Digital signal processors (DSP) are a specialized form of

microprocessor, while FPGAs are a form of highly configurable hardware.

Processing throughput

WiMAX and LTE broadband wireless systems have significantly higher throughput and data rate requirements than W-CDMA and cdma2000 cellular systems. To support these high data rates, the underlying hardware platform must have significant capability of processing throughput. In addition, advanced signal processing techniques such as Turbo coding/decoding, and front-end functions including fast Fourier transform/inverse fast Fourier transform (FFT/IFFT),beam-forming, MIMO, crest factor reduction (CFR), and digital pre-distortion (DPD) are computationally intensive and require several billion multiply and accumulate operations per second.

Flexibility

WiMAX is a relatively new market and is currently in the initial development and deployment stages. Similarly, 3GPP LTE is being defined and will go through numerous revisions before being finalized. While there are many competing mobile broadband technologies, such as WiMAX, LTE, and UMB, their common thread is OFDMA-MIMO (Parssinen et al., 2000). In this current scenario, having a flexible and reprogrammable product is necessary to provide a standards-agnostic or multi-protocol base station. Systems offering this flexibility can significantly reduce the capital expenditures and operating expenditures for wireless infrastructure OEMs and operators while alleviating risks posed by constantly evolving standards.

Cost-Reduction Path

A valuable lesson learned from designing and deploying 3G systems is the importance of establishing a long-term cost-reduction strategy in the beginning. Evolving WiMAX and LTE standards are expected to stabilize. For OEM sand service providers to remain competitive in the marketplace, the cost of the final product eventually will be more important than flexibility. Choosing the right hardware platform also provides a seamless cost-reduction path for production volumes, saving millions of dollars in engineering costs that would otherwise be incurred by system redesign.

#### **2.1.2 Generic hardware architecture**

The goal of LTE and WiMAX is to provide a high-data-rate, low-latency and packetoptimized radio-access technology supporting flexible bandwidth deployments. In addition, new network architecture is designed with the goal to support packet-switched traffic with seamless mobility, quality of service and minimal latency.

Due to the demand of high performance of high data-rate, low-latency and reduced delays, the choice of LTE hardware platform is a challenging job. The general architecture of the mobile communications system is depicted as the figure below.

The architecture can be divided into three parts: radio frequency (RF), intermediate frequency (IF) and digital baseband. The mainstream design of receiver proceed as follows: sampling the analog signal in the intermediate frequency, then down-converting the digital signal to baseband, finally demodulating the digital baseband signal. The transmitter is just opposite of the receiver (Dohler et al., 2005). Moreover, there may be CFR (Crest Factor Reduction) and DPD (Digital Pre-Distortion ) modules before converting to analog signal.

WiMAX and LTE broadband wireless systems have significantly higher throughput and data rate requirements than W-CDMA and cdma2000 cellular systems. To support these high data rates, the underlying hardware platform must have significant capability of processing throughput. In addition, advanced signal processing techniques such as Turbo coding/decoding, and front-end functions including fast Fourier transform/inverse fast Fourier transform (FFT/IFFT),beam-forming, MIMO, crest factor reduction (CFR), and digital pre-distortion (DPD) are computationally intensive and require several billion

WiMAX is a relatively new market and is currently in the initial development and deployment stages. Similarly, 3GPP LTE is being defined and will go through numerous revisions before being finalized. While there are many competing mobile broadband technologies, such as WiMAX, LTE, and UMB, their common thread is OFDMA-MIMO (Parssinen et al., 2000). In this current scenario, having a flexible and reprogrammable product is necessary to provide a standards-agnostic or multi-protocol base station. Systems offering this flexibility can significantly reduce the capital expenditures and operating expenditures for wireless infrastructure OEMs and operators while alleviating risks posed

A valuable lesson learned from designing and deploying 3G systems is the importance of establishing a long-term cost-reduction strategy in the beginning. Evolving WiMAX and LTE standards are expected to stabilize. For OEM sand service providers to remain competitive in the marketplace, the cost of the final product eventually will be more important than flexibility. Choosing the right hardware platform also provides a seamless cost-reduction path for production volumes, saving millions of dollars in engineering costs

The goal of LTE and WiMAX is to provide a high-data-rate, low-latency and packetoptimized radio-access technology supporting flexible bandwidth deployments. In addition, new network architecture is designed with the goal to support packet-switched traffic with

Due to the demand of high performance of high data-rate, low-latency and reduced delays, the choice of LTE hardware platform is a challenging job. The general architecture of the

The architecture can be divided into three parts: radio frequency (RF), intermediate frequency (IF) and digital baseband. The mainstream design of receiver proceed as follows: sampling the analog signal in the intermediate frequency, then down-converting the digital signal to baseband, finally demodulating the digital baseband signal. The transmitter is just opposite of the receiver (Dohler et al., 2005). Moreover, there may be CFR (Crest Factor Reduction) and DPD (Digital Pre-Distortion ) modules before converting to analog signal.

Processing throughput

Flexibility

multiply and accumulate operations per second.

that would otherwise be incurred by system redesign.

seamless mobility, quality of service and minimal latency.

mobile communications system is depicted as the figure below.

by constantly evolving standards.

**2.1.2 Generic hardware architecture** 

Cost-Reduction Path

Fig. 1. The general hardware architecture of mobile communications system

The process in IF contains NCO (numerical controlled oscillator), CIC filter, half-band filter and FIR filter. All of the modules mentioned above have such a simple structure that they can be easily implemented in an FPGA. Because of the high data-rate of IF digital data, it is impossible to implemented in DSP. For example, first decimation filter in a digital wireless receiver, typically, is a CIC filter, operating at a sample rate of 50-100MHz. At these rates any DSP processor would find it extremely difficult to do anything. However, the CIC has an extremely simple structure, and implementing it in an FPGA would be easy. A sample rate of 100MHz should be achievable, and even the smallest FPGA will have a lot of resource available for further processing. In addition, the latest mobile communications system has employed MIMO technology, which means there will be two or more antennas. As the fact of introducing MIMO, the parallel sampling data will be processed simultaneously. Therefore, it is general to choose FPGA or ASIC as the processor in IF instead of DSP. For example, AD6654 is an IF to baseband receiver, with programmable decimating FIR filters, interpolating half-band filters and CIC filters built-in (Fathi, 2004).

Baseband is usually divided into two parts. One of them is digital signal processing, which is used for implementing PHY protocol. The other part is microprocessor used for implementing MAC and lower protocol (Goldfarb et al., 2000). The microprocessor is generally selected within ARM or Power PC processor. However, the choice of baseband digital signal processing solution is various. Generally there are four choices which are ASIC, DSP, FPGA and DSP+FPGA (Jusslia et al., 2001). Application-specific integrated circuit (ASIC), is an integrated circuit (IC) customized for a particular use, rather than intended for general-purpose use. Digital signal processors (DSP) are a specialized form of microprocessor, while FPGAs are a form of highly configurable hardware.

Design and Implementation of WiMAX Baseband System 243

while the DSP could not. Equally, there are many complex software problems that the FPGA cannot address. Another advantage of co-processing is reconfigurable features of FPGA, which means that engineers can quickly build and modify the design architecture. Moreover, FPGA supports the integration of other components (such as Serial Rapid IO transceiver, PCI Express interfaces, glue logic and low-rate control task), which reduces overall system cost and power consumption. In addition, the integration of so many interfaces is valuable for scalability, which meets the changing demand of mobile communications system. As a result, the ideal system is often to split the work between

In the following discussion, we adopt the picoChip PC7205 development platform as our baseband hardware platform, in which integrated one of the multi-core DSP processor (PC205) and one piece of FPGA. The PC205 process also includes an embedded ARM926EJ

The PC205 microprocessor interface is designed for communications with a processor. No specific processor family is assumed and data can be exchanged over 8, 16 or 32-bit wide data bus. The processor interface is used for communication between MAC and PHY.

2. Burst (DMA) – Reading or writing words at the same rate as the microprocessor proc

GPR accesses allow access to the majority of memory mapped registers and services within the processor interface, GPR accesses can only be used for single read / write accesses. Typically, GPR is used for transmit control signals whose amount of data is small such as

FPGAs and DSPs.

Fig. 2. PC205 block diagram

clock.

**2.1.3 Communication between MAC and PHY** 

processor that could implement the MAC functions.

The Processor interface supports two basic transaction types 1. Single (GPR) – Reading or writing one word at a time.

automatic power control (APC) signal between MAC and PHY.

#### ASIC

According to circuit functions and performance requirements, ASIC design needs to select circuit form, the device structure, process plan and design rules to minimize chip area, lower design cost and shorten the design cycle, and finally brings forward the correct and reasonable mask layout. Nevertheless, the disadvantages of full-custom design can include increased manufacturing and design time, increased non-recurring engineering costs, more complexity in the computer-aided design (CAD) system. Moreover, Due to the changing demand of mobile communications system, the equipment has to upgrade one day, but ASIC cannot upgrade flexibly. When the hardware platform does not meet the requirements, all of the equipment must be replaced. As a result, the cost of upgrading is very expensive.

DSP

A digital signal processor (DSP) is a specialized microprocessor with an optimized architecture for the fast operational needs of digital signal processing. Digital signal processing algorithms typically require a large number of mathematical operations to be performed quickly and repetitively on a set of data. Signals (perhaps from audio or video sensors) are constantly converted from analog to digital, manipulated digitally, and then converted again to analog form. Many DSP applications have constraints on latency; that is, for the system to work, the DSP operation must be completed within some fixed time, and deferred (or batch) processing is not viable (Parssinen et al, 1999).

Multi-core DSP

Multi-core processing is the technology or group of technologies that companies like Intel and IBM are betting will replace Instruction Level Parallelism and the clock rate ratchet: dual and quad core systems for desktop applications are already in volume production. As we have seen, in many cases the controlling factor in device performance has moved from the ability to complete computation to the ability to move data. Well designed multi-core architectures allow data stores from registers to main memory to be distributed throughout the system, in whatever way makes most sense for the application. In fact, in multi-core architectures the communications fabric can substitute for memory accesses by allowing direct communication between the processing elements. If matched to the task in hand, such an infrastructure can therefore intrinsically help to overcome any restrictions imposed by the need to move data.

FPGA

The FPGA configuration is generally specified using a hardware description language (HDL), similar to that used for an application-specific integrated circuit (ASIC) (circuit diagrams were previously used to specify the configuration, as they were for ASICs, but this is increasingly rare). FPGAs can be used to implement any logical function that an ASIC could perform. The ability to update the functionality after shipping, partial reconfiguration of the portion of the design and the low non-recurring engineering costs relative to an ASIC design (notwithstanding the generally higher unit cost) offer advantages for many applications.

FPGA-DSP Co-Processing

FPGA and DSP represent two very different approaches to signal processing – each good at different things. There are many high sampling rate applications that an FPGA does easily,

According to circuit functions and performance requirements, ASIC design needs to select circuit form, the device structure, process plan and design rules to minimize chip area, lower design cost and shorten the design cycle, and finally brings forward the correct and reasonable mask layout. Nevertheless, the disadvantages of full-custom design can include increased manufacturing and design time, increased non-recurring engineering costs, more complexity in the computer-aided design (CAD) system. Moreover, Due to the changing demand of mobile communications system, the equipment has to upgrade one day, but ASIC cannot upgrade flexibly. When the hardware platform does not meet the requirements, all of the

A digital signal processor (DSP) is a specialized microprocessor with an optimized architecture for the fast operational needs of digital signal processing. Digital signal processing algorithms typically require a large number of mathematical operations to be performed quickly and repetitively on a set of data. Signals (perhaps from audio or video sensors) are constantly converted from analog to digital, manipulated digitally, and then converted again to analog form. Many DSP applications have constraints on latency; that is, for the system to work, the DSP operation must be completed within some fixed time, and

Multi-core processing is the technology or group of technologies that companies like Intel and IBM are betting will replace Instruction Level Parallelism and the clock rate ratchet: dual and quad core systems for desktop applications are already in volume production. As we have seen, in many cases the controlling factor in device performance has moved from the ability to complete computation to the ability to move data. Well designed multi-core architectures allow data stores from registers to main memory to be distributed throughout the system, in whatever way makes most sense for the application. In fact, in multi-core architectures the communications fabric can substitute for memory accesses by allowing direct communication between the processing elements. If matched to the task in hand, such an infrastructure can therefore intrinsically help to overcome any restrictions imposed by

The FPGA configuration is generally specified using a hardware description language (HDL), similar to that used for an application-specific integrated circuit (ASIC) (circuit diagrams were previously used to specify the configuration, as they were for ASICs, but this is increasingly rare). FPGAs can be used to implement any logical function that an ASIC could perform. The ability to update the functionality after shipping, partial reconfiguration of the portion of the design and the low non-recurring engineering costs relative to an ASIC design (notwithstanding the generally higher unit cost) offer advantages

FPGA and DSP represent two very different approaches to signal processing – each good at different things. There are many high sampling rate applications that an FPGA does easily,

equipment must be replaced. As a result, the cost of upgrading is very expensive.

deferred (or batch) processing is not viable (Parssinen et al, 1999).

ASIC

DSP

Multi-core DSP

the need to move data.

for many applications.

FPGA-DSP Co-Processing

FPGA

while the DSP could not. Equally, there are many complex software problems that the FPGA cannot address. Another advantage of co-processing is reconfigurable features of FPGA, which means that engineers can quickly build and modify the design architecture. Moreover, FPGA supports the integration of other components (such as Serial Rapid IO transceiver, PCI Express interfaces, glue logic and low-rate control task), which reduces overall system cost and power consumption. In addition, the integration of so many interfaces is valuable for scalability, which meets the changing demand of mobile communications system. As a result, the ideal system is often to split the work between FPGAs and DSPs.

#### **2.1.3 Communication between MAC and PHY**

In the following discussion, we adopt the picoChip PC7205 development platform as our baseband hardware platform, in which integrated one of the multi-core DSP processor (PC205) and one piece of FPGA. The PC205 process also includes an embedded ARM926EJ processor that could implement the MAC functions.

Fig. 2. PC205 block diagram

The PC205 microprocessor interface is designed for communications with a processor. No specific processor family is assumed and data can be exchanged over 8, 16 or 32-bit wide data bus. The processor interface is used for communication between MAC and PHY.

The Processor interface supports two basic transaction types


GPR accesses allow access to the majority of memory mapped registers and services within the processor interface, GPR accesses can only be used for single read / write accesses. Typically, GPR is used for transmit control signals whose amount of data is small such as automatic power control (APC) signal between MAC and PHY.

Design and Implementation of WiMAX Baseband System 245

Generally there are two kinds of interface signals between baseband and RF, which are data and control signals respectively. Data signal are usually 16 bits complex format. However, the control signals may adopt one communication protocol such as SPI, I2C and so on. Moreover, some of the protocols may be changed for the sake of implementation. The interface of DSP may not support all the protocols. As the result, it is the right way to introduce one piece of FPGA between the baseband and RF for flexibility and scalability. We

In addition, for the sake of power consumption, more and more DSPs have chosen 1V and 1.8V as power supply of the core and interfaces respectively. But the other device may take 3.3V as the power supply of the interfaces. It is obvious that electrical characteristics don`t matched between the different interfaces. Nowadays, most of FPGA have more than one bank, and each bank can be supplied different power. We can use some of banks with 1.8V power as the interfaces with DSP, while the other banks with 3.3V power as the interfaces

Automatic gain control (AGC) Automatic gain control (AGC) is an adaptive function found in many electronic devices. In a digital communication receiver strong signals that fall outside the narrowband digital filter bandwidth, but inside the analog IF translator bandwidths, can overload or saturate the A/D converter. This results in the generation of in-band IMD products and can result in significant degradation of the desired signal. If large signal levels are detected at the A/D converter, the receiver gain may have to be redistributed by reducing the pre-conversion analog gain and increasing the digital gain to maintain the desired signal output level. This will, however, reduce the desired signal-to-

In the WiMAX system, there are two mechanisms for power control, which are open loop power control and closed loop power control respectively. It is necessary to have closed loop power control while open power control is optional. Closed loop power control means that the base station (BS) controls the transmission power of the mobile station (MS). The MS transmission power is controlled in order to avoid exceeding the BS`s total receiving power from an antenna. In the WiMAX standard, other uses of it are not defined (i.e., the uplink

Usually RF device has the specific module for receiving gain and transmission power adjusting, which is controlled by voltage signals. Therefore, baseband just outputs direct current signals with variable amplitude to RF. Typically we can obtain the direct current signals with low rate DAC, but the interface between baseband and RF have to increases parallel lines used for transmitting the digital. Here we introduce a simple method to

**2.2 Connecting baseband to RF** 

quantization noise ratio.

TPC algorithm is vendor specific).

generate the direct current signals.

AGC and Power Control signal design

Power Control

**2.2.1 Connecting to RF by using FPGA** 

can write the suitable protocols for almost all interfaces in FPGA.

with some other device. As a result, it is easy to change the logic level.

**2.2.2 Automatic gain control (AGC) and power control** 

DMA Accesses are primarily used for the efficient movement of data to and from the picoArray. Generally we use DMA to transmit the bulk data such as wireless frame between MAC and PHY. Fig.3 shows the DMA channel configured for write access. A FIFO buffers the data written from the microprocessor to the selected DMA channel. The FIFO output is connected to the picoBus and data is transferred to the internal array elements by using a get command from within the software.

Fig. 3. DMA channel configured for write access

The PC205 support 3 DMA transfer mechanisms.


In the mode of Basic downlink, host processor initiates transfer. The process is as followed.


In the mode of basic uplink, host processor initiates transfer, but the processor has to preassumes data size, which is not practical.

In the mode of HWIF\_UL, PicoArray uses handshake mechanism to indicate data size through GPR registers, and uses interrupt to initiate transfer through ITS register. The process is as followed.


#### **2.2 Connecting baseband to RF**

244 Advanced Transmission Techniques in WiMAX

DMA Accesses are primarily used for the efficient movement of data to and from the picoArray. Generally we use DMA to transmit the bulk data such as wireless frame between MAC and PHY. Fig.3 shows the DMA channel configured for write access. A FIFO buffers the data written from the microprocessor to the selected DMA channel. The FIFO output is connected to the picoBus and data is transferred to the internal array elements by using a

get command from within the software.

Fig. 3. DMA channel configured for write access The PC205 support 3 DMA transfer mechanisms. 1. Basic downlink Host processor to picoArray

3. HWIF\_UL picoArray to Host processor

 Open a transport session Configure the transport session Start the transport session

Close the transport session

assumes data size, which is not practical.

Poll for an event indicating the date available.

Write DMA data

process is as followed.

interrupts.

Read DMA data

Close the transport session

2. Basic uplink picoArray to Host processor – not practical

In the mode of Basic downlink, host processor initiates transfer. The process is as followed.

In the mode of basic uplink, host processor initiates transfer, but the processor has to pre-

In the mode of HWIF\_UL, PicoArray uses handshake mechanism to indicate data size through GPR registers, and uses interrupt to initiate transfer through ITS register. The

Open and setup an interrupt clearing transport session used for clearing down

Open and setup a HWIF\_UL transport session for uplink DMA.

#### **2.2.1 Connecting to RF by using FPGA**

Generally there are two kinds of interface signals between baseband and RF, which are data and control signals respectively. Data signal are usually 16 bits complex format. However, the control signals may adopt one communication protocol such as SPI, I2C and so on. Moreover, some of the protocols may be changed for the sake of implementation. The interface of DSP may not support all the protocols. As the result, it is the right way to introduce one piece of FPGA between the baseband and RF for flexibility and scalability. We can write the suitable protocols for almost all interfaces in FPGA.

In addition, for the sake of power consumption, more and more DSPs have chosen 1V and 1.8V as power supply of the core and interfaces respectively. But the other device may take 3.3V as the power supply of the interfaces. It is obvious that electrical characteristics don`t matched between the different interfaces. Nowadays, most of FPGA have more than one bank, and each bank can be supplied different power. We can use some of banks with 1.8V power as the interfaces with DSP, while the other banks with 3.3V power as the interfaces with some other device. As a result, it is easy to change the logic level.

#### **2.2.2 Automatic gain control (AGC) and power control**

Automatic gain control (AGC) Automatic gain control (AGC) is an adaptive function found in many electronic devices. In a digital communication receiver strong signals that fall outside the narrowband digital filter bandwidth, but inside the analog IF translator bandwidths, can overload or saturate the A/D converter. This results in the generation of in-band IMD products and can result in significant degradation of the desired signal. If large signal levels are detected at the A/D converter, the receiver gain may have to be redistributed by reducing the pre-conversion analog gain and increasing the digital gain to maintain the desired signal output level. This will, however, reduce the desired signal-toquantization noise ratio.

Power Control

In the WiMAX system, there are two mechanisms for power control, which are open loop power control and closed loop power control respectively. It is necessary to have closed loop power control while open power control is optional. Closed loop power control means that the base station (BS) controls the transmission power of the mobile station (MS). The MS transmission power is controlled in order to avoid exceeding the BS`s total receiving power from an antenna. In the WiMAX standard, other uses of it are not defined (i.e., the uplink TPC algorithm is vendor specific).

AGC and Power Control signal design

Usually RF device has the specific module for receiving gain and transmission power adjusting, which is controlled by voltage signals. Therefore, baseband just outputs direct current signals with variable amplitude to RF. Typically we can obtain the direct current signals with low rate DAC, but the interface between baseband and RF have to increases parallel lines used for transmitting the digital. Here we introduce a simple method to generate the direct current signals.

Design and Implementation of WiMAX Baseband System 247

This block provides the main data path for transmit in the PHY, which comprises the

Unpacking of data into FEC blocks, Encoding, including randomization, FEC,

Data modulation: The conversion from bit stream to QPSK, 16QAM, 64QAM symbols---

 OFDMA zone processing, including pilot and preamble generation subcarrier permutation, subcarrier scrambling and zone boost---BurstZoneblock&AntEnc block Antenna processing, including IFFT, cyclic prefix, Peak-to Average reduction and

This block provides the main data path for receive in the PHY, which including:

Time and frequency synchronization process---MsRxAcq block &MsRxTf block

Antenna processing with receiving filtering (Foschini, 1996), ALC, cyclic prefix removal

Fig. 4. Main function blocks of PHY Layer

interleaving, repetition---FEC block

transmitting filtering---FrondEnd block

Transmit-path:

following stages:

ConstPack block

Fig. 5. BS transmit process

and FFT---MsRxTf block

Receive-path:

We can make use of Pulse Width Modulation (PWM) signal and a RC low-pass filter to generate the direct current signal. When the PWM signal duty ratio is 100 percent, the amplitude of direct current signal equals to the amplitude of PWM signal. When the PWM signal duty ratio is 50 percent, the amplitude of direct current signal equals to the half amplitude of PWM signal. The direct current signal is approximately linear with the duty cycle. In this case, it is necessary to use two digital signal lines for receiving gain and transmission power control.

#### **2.2.3 Extern GPS synchronization signal**

The IEEE 802.16 standard calls for the use of global positioning system (GPS) receivers to provide the precise time reference for synchronization of WiMAX networks. This operation is performed both during the startup and periodically in order to maintain the alignment with the external PPS pulse.

Briefly, the algorithm follows these steps. The controller of synchronization starts searching for the first PPS pulse while discarding the RX samples. Then it stalls the PHY while waiting for the PPS pulse and sends DL dummy complex samples. Once received the PPS pulse, after 100 ms, the controller starts passing the DL complex samples. For each frame period, the frame synchronization module receives the frame start indication and decides when a frame adjustment is required for maintaining the alignment with the external pulse.

#### **3. Physical layer implementation**

#### **3.1 Introduction of PHY**

Considering PHY implementation, the main functions of PHY layer may comprise API, control, transmit-path, receive-path and synchronization/radio interface for both BS and MS entities (LAN/MAN Standards Committee of the IEEE Computer Society et al., 2008). Fig.4depicts the relationship between these function blocks.

#### API:

The API provides an interface between the PHY and MAC. Its function is responsible for:


#### Control:

The control function is responsible for distributing control data originating in the API around the other functions of the PHY. Its most important task is to ensure that each of the data-path functions has access to the control data that it needs to process the data-path data it is currently working on. Secondly it orchestrates the collection of measurements from the PHY and provides routes for diagnostic information to be accessed by the MAC or other external processes.

Fig. 4. Main function blocks of PHY Layer

Transmit-path:

246 Advanced Transmission Techniques in WiMAX

We can make use of Pulse Width Modulation (PWM) signal and a RC low-pass filter to generate the direct current signal. When the PWM signal duty ratio is 100 percent, the amplitude of direct current signal equals to the amplitude of PWM signal. When the PWM signal duty ratio is 50 percent, the amplitude of direct current signal equals to the half amplitude of PWM signal. The direct current signal is approximately linear with the duty cycle. In this case, it is necessary to use two digital signal lines for receiving gain and

The IEEE 802.16 standard calls for the use of global positioning system (GPS) receivers to provide the precise time reference for synchronization of WiMAX networks. This operation is performed both during the startup and periodically in order to maintain the alignment

Briefly, the algorithm follows these steps. The controller of synchronization starts searching for the first PPS pulse while discarding the RX samples. Then it stalls the PHY while waiting for the PPS pulse and sends DL dummy complex samples. Once received the PPS pulse, after 100 ms, the controller starts passing the DL complex samples. For each frame period, the frame synchronization module receives the frame start indication and decides when a

Considering PHY implementation, the main functions of PHY layer may comprise API, control, transmit-path, receive-path and synchronization/radio interface for both BS and MS entities (LAN/MAN Standards Committee of the IEEE Computer Society et al., 2008).

The API provides an interface between the PHY and MAC. Its function is responsible for:

Interpreting data from the MAC and generating internal control data for other functions

The control function is responsible for distributing control data originating in the API around the other functions of the PHY. Its most important task is to ensure that each of the data-path functions has access to the control data that it needs to process the data-path data it is currently working on. Secondly it orchestrates the collection of measurements from the PHY and provides routes for diagnostic information to be accessed by the MAC or other

frame adjustment is required for maintaining the alignment with the external pulse.

transmission power control.

with the external PPS pulse.

**3.1 Introduction of PHY** 

in the PHY

external processes.

Control:

API:

**2.2.3 Extern GPS synchronization signal** 

**3. Physical layer implementation** 

Fig.4depicts the relationship between these function blocks.

The physical transfer of data to and from the MAC

Interpreting data from the PHY and parsing into data for the MAC

 Buffering of data to and from the MAC Error checking and diagnostics on the data This block provides the main data path for transmit in the PHY, which comprises the following stages:


Fig. 5. BS transmit process

Receive-path:

This block provides the main data path for receive in the PHY, which including:


Design and Implementation of WiMAX Baseband System 249

The MATLAB simulation platform is built in the way that all blocks of the platform are map to the functions on picoArray respectively. So, the function blocks are transferred from MATLAB platform to DSP platform smoothly. Moreover, each block of the MATLAB fixedpoint simulation can generate the corresponding result of this block which can be used as input of the followed block on the DSP platform directly. That is a very efficient way to

Matlab platform is built in accordance with the WiMAX physical layer protocol. The platform complies with the frame structure and resource distribution of WiMAX standard. It can generate any structure's sub-frame. Also it can provide fixed-point simulation for each sub-frame. Moreover, the platform can generate test vectors for each of the WiMAX PHY's module. The test vectors can be mapped to the AE level (picoChip processing unit),

In the following, the PHY protocol implementation on picoArray DSP, the chosen Multi-

The picoArray multi-core DSP is based on a massively parallel architecture comprising large numbers of small independent processors. A DSP application is logically decomposed into a number of communicating sequential processes, each of which is assigned to a particular processor on the picoArray. The designing tools statically allocate processors and picoBus (PicoChip Company, 2008) resources for the system, so there is no need for an operating system. The static allocation of resources allows much of the system's runtime complexity to be moved back into the tool suite. It allows the hardware to be lightweight and hence allows a very high proportion of the power of the processor array to be used for the real DSP

Fig.7 shows a simplified representation of the overall structure of the picoArray. Each box marked 'P' in the figure represents a single processor, referred to an Array Element(AE).

including the input control information and the input and output data of each AE.

**3.2.2 Simulation platform development** 

verify the functions on the picoArray.

**3.3.1 PicoArray introduction** 

application.

**3.3 PHY protocol implementation on picoArray DSP** 

Fig. 7. A simplified representation of a picoArray

Core DSP for baseband application will be introduced in details.


Fig. 6. MS receive process

Synchronization /radio interface:

This block is responsible for controlling the absolute and relative timing of uplink and downlink frames in the PHY and the radio. It is also responsible formultiplexing, demultiplexing and formatting data for the interface to the radio via the picoArray ADI (Asynchronous Data Interface) interface(PicoChip Company, 2008). For this reason realizing block may well be somewhat platform specific.

#### **3.2 Link-level simulation**

#### **3.2.1 Simulation platform based on MATLAB**

Before realizing the whole WiMAX PHY layer software on the picoArray DSP (PicoChip Company, 2008), a fixed-point link-level simulation is needed. First of all, it's important to make sure that the algorithms are correct and satisfy the performance demand in simulation environment. Because the development on DSP processor is time-consuming and expensive, the consequence of implementing a system that will never work in DSP processor can't be affordable. Secondly, it's very difficult to locate bugs and correct them in DSP processor. When the bugs have nothing to do with hardware, the bugs finding and correcting work can move back to simulation platform. This will save your development time and cost significantly. At last, there are varies wireless channel models in MATLAB, which are very useful for us to figure out how the system performs on different channel environment.

MATLAB simulation platforms are floating point in common situations. But this simulation platform does an extra job that it converts the calculation result from floating point to fixedpoint result. That is to say, the simulation platform is a fixed-point platform which can be more approaching to fixed-point picoArray DSP processor. In this way, the performance between MATLAB simulation platform and DSP process on picoArray will be the same roughly. It makes the simulation more convincing.

#### **3.2.2 Simulation platform development**

248 Advanced Transmission Techniques in WiMAX

 OFDMA zone receive, with AAS/MIMO (Mugenet al., 2007) processing and buffering, subcarrier descrambling and depermutation, pilot extraction, CPE and frequency compensation, channel estimation and equalization, constellation demaping, channel

 Decoding: including derepetition, deinterleaving, depuncturing, H-ARQ (Lin and Yu, July 1982), FEC decoding, derandomising and repacking of user data into PDUs---

This block is responsible for controlling the absolute and relative timing of uplink and downlink frames in the PHY and the radio. It is also responsible formultiplexing, demultiplexing and formatting data for the interface to the radio via the picoArray ADI (Asynchronous Data Interface) interface(PicoChip Company, 2008). For this reason realizing

Before realizing the whole WiMAX PHY layer software on the picoArray DSP (PicoChip Company, 2008), a fixed-point link-level simulation is needed. First of all, it's important to make sure that the algorithms are correct and satisfy the performance demand in simulation environment. Because the development on DSP processor is time-consuming and expensive, the consequence of implementing a system that will never work in DSP processor can't be affordable. Secondly, it's very difficult to locate bugs and correct them in DSP processor. When the bugs have nothing to do with hardware, the bugs finding and correcting work can move back to simulation platform. This will save your development time and cost significantly. At last, there are varies wireless channel models in MATLAB, which are very useful for us to figure out how the system performs on different channel environment.

MATLAB simulation platforms are floating point in common situations. But this simulation platform does an extra job that it converts the calculation result from floating point to fixedpoint result. That is to say, the simulation platform is a fixed-point platform which can be more approaching to fixed-point picoArray DSP processor. In this way, the performance between MATLAB simulation platform and DSP process on picoArray will be the same

state compensation and MRC---MsRxSym block &MsRxMap block

MsRxBurstChain block

Fig. 6. MS receive process

**3.2 Link-level simulation** 

Synchronization /radio interface:

block may well be somewhat platform specific.

**3.2.1 Simulation platform based on MATLAB** 

roughly. It makes the simulation more convincing.

The MATLAB simulation platform is built in the way that all blocks of the platform are map to the functions on picoArray respectively. So, the function blocks are transferred from MATLAB platform to DSP platform smoothly. Moreover, each block of the MATLAB fixedpoint simulation can generate the corresponding result of this block which can be used as input of the followed block on the DSP platform directly. That is a very efficient way to verify the functions on the picoArray.

Matlab platform is built in accordance with the WiMAX physical layer protocol. The platform complies with the frame structure and resource distribution of WiMAX standard. It can generate any structure's sub-frame. Also it can provide fixed-point simulation for each sub-frame. Moreover, the platform can generate test vectors for each of the WiMAX PHY's module. The test vectors can be mapped to the AE level (picoChip processing unit), including the input control information and the input and output data of each AE.

#### **3.3 PHY protocol implementation on picoArray DSP**

In the following, the PHY protocol implementation on picoArray DSP, the chosen Multi-Core DSP for baseband application will be introduced in details.

#### **3.3.1 PicoArray introduction**

The picoArray multi-core DSP is based on a massively parallel architecture comprising large numbers of small independent processors. A DSP application is logically decomposed into a number of communicating sequential processes, each of which is assigned to a particular processor on the picoArray. The designing tools statically allocate processors and picoBus (PicoChip Company, 2008) resources for the system, so there is no need for an operating system. The static allocation of resources allows much of the system's runtime complexity to be moved back into the tool suite. It allows the hardware to be lightweight and hence allows a very high proportion of the power of the processor array to be used for the real DSP application.

Fig. 7. A simplified representation of a picoArray

Fig.7 shows a simplified representation of the overall structure of the picoArray. Each box marked 'P' in the figure represents a single processor, referred to an Array Element(AE).

Design and Implementation of WiMAX Baseband System 251

After developers have selected AEs for each block, memory access type should be selected. There are three types of data memory for data storage: data registers inside AE, data memory inside AE and SDRAM outside AE. The data registers are very fast access memories. When storing a small amount of variables, developers ought to use the registers as far as possible. There are only 15 registers in one AE as shown in table 1. If the size of data which needs to be stored exceeds the amount of unused registers, the data should be stored in the inside data memory. The access speed of this type of memory is a little slower than registers but is much faster than SDRAM outside of AEs. Last but not least, if the size of data exceeds the capacity of the memory inside AE, SDRAM is used to store it. In this situation, developers must arrange the SDRAM access area very carefully because the SDRAM is shared among different AEs which need it for their data storage. If some of the AEs use the same area in the SDRAM, fatal error will take place unexpectedly and is

Then it comes to programming step. Firstly, functional code is created carefully based on the MATLAB simulation platform for each block. Then verify the code's syntax accuracy and logical accuracy with picoTools. The next step is taking the MATLAB simulation test vector as the input vector of each block. The function of block on every AE is verified by through comparison between the AE output and the MATLAB simulation result. The throughput matching is another important process when programming. The reason is if the throughputs among the blocks don't fit for each other, they can't work when connected together. If this problem happens, you should change your design to matching the throughputs demand. One principle is that the getting/reading port rate of the AE must be faster than the rate of

Last but not least, debug on picoArray. All the above coding and debug work is done with picoTools on development environment. It is easy to select the result of each block in the form of text document to be compared with MATLAB simulation result for verification. When it comes to the debug work on the picoArray, it is much more difficult to get the result of each block. As some hardware-related bugs can't be discovered on software environment, it's very important to get some methods for the debug work on picoArray. Fortunately, a probe mechanism is provided. You can configure the unused AE or unused SDRAM to get the output of the block which is needed to debug. The data in the 'probe' AE or SDRAM can be transmitted into text document to compare with MATLAB simulation

First, the sum of the rates of signals connected to one single AE can't exceed 2 because there are two channels connecting one AE to picoBus. For example, if one AE has three signals names sig\_A, sig\_B and sig\_C. The rates of the three signals are @2, @2 and @1 respectively. It is easy to find out that 1/2+1/2+1/1=2, so there are no channel space for another signal to

Secondly, for a process chain, the getting data rate of AE must be faster than the providing data rate of the previous AE. This is very important rule during the picoArray DSP design, because if this can't be satisfied, data flow would be blocked. Moreover, if one AE's data

difficult to discover.

the AE putting/writing data.

result of the same block.

connect to this AE.

Some tips of developing on the picoArrays are given as follows:

flow is blocked, the conjoint AE' data flow will be blocked too.

The processors are laid out in a grid, interconnected by a matrix of buses called the picoBus. Each AE is connected to two buses. The lines between the processors represent the picoBus, and the circles represent bus switches which connect buses together to provide routes between all AEs in the array. The communications between these processes, called signals, are then mapped on to physical segments of the picoBus between the assigned processors by suitable settings of the bus switches. The heavy red and blue lines illustrate two example connection paths between particular processors. Communication between AEs is timemultiplexed over the picoBus, which is a shared resource.

#### **3.3.2 PHY implementation**

A PicoArray process is composed of a number of AE which can work simultaneously. This structure is quite different from traditional single-core processor. As a result, developing work on picoArray will share nothing with that on traditional single-core processor. The major steps of developing work on picoArray are given as follows.

All the Instruction/Data memory that developers can utilized are in the core in traditional processors, so developers needn't care about how to arrange Instruction/Data memory for each functional block. But when doing developing work on picoArray, the first thing developers need to deal with is to select suitable AE for each functional block. There are some different categories of AE which have quite different abilities. Three types of AE are mostly used in our work: STAN2 (short for standard AE), MEM2 (short for memory AE) and CTRL2 (short for control AE).


Table 1. AE types in PicoArray processor

For example, there are 196 STAN2 AEs, 50 MEM2 AEs and 2 CTRL AEs in one piece of picoArray PC205. The ability contrast is shown in the above table. When arranging the functional blocks into AEs, developers should select a good enough type of AE to implement the block according to the need of the blocks. Moreover, developers should optimize the code of the block to fulfill the limit of different types of AE. If one piece of picoArray can't hold all the blocks, another piece of picoArray processor should be introduced in to share the burden. The communication between the two pieces of picoArray processor is accomplished by IPI (Inter PicoArray Interface).

The processors are laid out in a grid, interconnected by a matrix of buses called the picoBus. Each AE is connected to two buses. The lines between the processors represent the picoBus, and the circles represent bus switches which connect buses together to provide routes between all AEs in the array. The communications between these processes, called signals, are then mapped on to physical segments of the picoBus between the assigned processors by suitable settings of the bus switches. The heavy red and blue lines illustrate two example connection paths between particular processors. Communication between AEs is time-

A PicoArray process is composed of a number of AE which can work simultaneously. This structure is quite different from traditional single-core processor. As a result, developing work on picoArray will share nothing with that on traditional single-core processor. The

All the Instruction/Data memory that developers can utilized are in the core in traditional processors, so developers needn't care about how to arrange Instruction/Data memory for each functional block. But when doing developing work on picoArray, the first thing developers need to deal with is to select suitable AE for each functional block. There are some different categories of AE which have quite different abilities. Three types of AE are mostly used in our work: STAN2 (short for standard AE), MEM2 (short for memory AE)

Feature CTRL2 MEM2 STAN2

Bus connections 4 2 2 Number of ports 32 12 10

> 49152/16384232768/327681 16384/49152 0 default 0

registers 15 15 15

For example, there are 196 STAN2 AEs, 50 MEM2 AEs and 2 CTRL AEs in one piece of picoArray PC205. The ability contrast is shown in the above table. When arranging the functional blocks into AEs, developers should select a good enough type of AE to implement the block according to the need of the blocks. Moreover, developers should optimize the code of the block to fulfill the limit of different types of AE. If one piece of picoArray can't hold all the blocks, another piece of picoArray processor should be introduced in to share the burden. The communication between the two pieces of picoArray

6656/2048 3 4608/4096 2 2560/6144 1 512/8192 0 default 0

Yes Yes No

512/256

multiplexed over the picoBus, which is a shared resource.

major steps of developing work on picoArray are given as follows.

**3.3.2 PHY implementation** 

and CTRL2 (short for control AE).

Number of

Instruction/Data memory size options (bytes)

> Byte memory accesses

Table 1. AE types in PicoArray processor

processor is accomplished by IPI (Inter PicoArray Interface).

After developers have selected AEs for each block, memory access type should be selected. There are three types of data memory for data storage: data registers inside AE, data memory inside AE and SDRAM outside AE. The data registers are very fast access memories. When storing a small amount of variables, developers ought to use the registers as far as possible. There are only 15 registers in one AE as shown in table 1. If the size of data which needs to be stored exceeds the amount of unused registers, the data should be stored in the inside data memory. The access speed of this type of memory is a little slower than registers but is much faster than SDRAM outside of AEs. Last but not least, if the size of data exceeds the capacity of the memory inside AE, SDRAM is used to store it. In this situation, developers must arrange the SDRAM access area very carefully because the SDRAM is shared among different AEs which need it for their data storage. If some of the AEs use the same area in the SDRAM, fatal error will take place unexpectedly and is difficult to discover.

Then it comes to programming step. Firstly, functional code is created carefully based on the MATLAB simulation platform for each block. Then verify the code's syntax accuracy and logical accuracy with picoTools. The next step is taking the MATLAB simulation test vector as the input vector of each block. The function of block on every AE is verified by through comparison between the AE output and the MATLAB simulation result. The throughput matching is another important process when programming. The reason is if the throughputs among the blocks don't fit for each other, they can't work when connected together. If this problem happens, you should change your design to matching the throughputs demand. One principle is that the getting/reading port rate of the AE must be faster than the rate of the AE putting/writing data.

Last but not least, debug on picoArray. All the above coding and debug work is done with picoTools on development environment. It is easy to select the result of each block in the form of text document to be compared with MATLAB simulation result for verification. When it comes to the debug work on the picoArray, it is much more difficult to get the result of each block. As some hardware-related bugs can't be discovered on software environment, it's very important to get some methods for the debug work on picoArray. Fortunately, a probe mechanism is provided. You can configure the unused AE or unused SDRAM to get the output of the block which is needed to debug. The data in the 'probe' AE or SDRAM can be transmitted into text document to compare with MATLAB simulation result of the same block.

Some tips of developing on the picoArrays are given as follows:

First, the sum of the rates of signals connected to one single AE can't exceed 2 because there are two channels connecting one AE to picoBus. For example, if one AE has three signals names sig\_A, sig\_B and sig\_C. The rates of the three signals are @2, @2 and @1 respectively. It is easy to find out that 1/2+1/2+1/1=2, so there are no channel space for another signal to connect to this AE.

Secondly, for a process chain, the getting data rate of AE must be faster than the providing data rate of the previous AE. This is very important rule during the picoArray DSP design, because if this can't be satisfied, data flow would be blocked. Moreover, if one AE's data flow is blocked, the conjoint AE' data flow will be blocked too.

Design and Implementation of WiMAX Baseband System 253

According to the function, the API is divided into two parts: communication with the MAC and the data processing chain of PHY layer. The first part includes getting messages from MAC, sending data to the SDRAM and regrouping the receive message to MAC. The messages from MAC include control-plane messages, which are used to perform configuration and reconfiguration of the PHY, and data-plane messages. Both messages have the same message format which includes message header, which descript message type and PHY entity, and message body. The receive messages to MAC are called response or indication messages. Then it comes to communication with PHY. The input signals of each module in the data processing chain can be classified into two types: control and data signal. Among them, all control signals come from the API; data signal is the output of the previous module. However, the beginning of the data also derives from the API. So, API separately process control and data information to produce the two signals. Their modules

According to the OSI seven-layer network protocol,MAC lays between the PHY layer and the network layer, responsibility for the data convergence and resource scheduling. So there

are Control system and Data processing, which are shown in Fig. 9.

**3.4.2 The block diagram of API** 

Fig. 9. The block diagram of API

**4. MAC protocol implementation 4.1 Introduction of MAC functions** 

#### **3.4 API architecture design**

#### **3.4.1 API introduction**

The API interface is between the MAC and OFDMA PHY, which is defined to picoChip's IEEE 802.16e base-station PHY and optional lower MAC accelerator to perform CRC and HCS calculations. The API described in this document is based on the Wireless MAN-OFDMA PHY. The greatest feature of API is to process data and control information separately.


Fig. 8. 802.16e Protocol Stack

Fig.8 shows a modified version of the 802.16e reference model and the location of the API as described in this document (shown as MAC/PHY API). The 802.16e protocol stack is shown, together with the radio and management entity.

#### **3.4.2 The block diagram of API**

252 Advanced Transmission Techniques in WiMAX

The API interface is between the MAC and OFDMA PHY, which is defined to picoChip's IEEE 802.16e base-station PHY and optional lower MAC accelerator to perform CRC and HCS calculations. The API described in this document is based on the Wireless MAN-OFDMA PHY. The greatest feature of API is to process data and control information

 The API addresses data and control plane functions. Data plane functions include the transfer of MAC PDUs in the uplink and downlink directions via the PHY service access point (SAP). Information required for uplink and downlink processing is sent

 Control-plane functions include determining the capabilities of the PHY, reconfiguration of the PHY and notification of error conditions, alarms and

 The base-station PHY can be configured to perform CRC and HCS calculations to lessen the processor requirements for the MAC. Encryption and decryption is beyond the

A frame-sync interrupt is sent to indicate the start of every downlink sub-frame; this

 The response and indication primitives sent from the PHY to MAC can be masked at PHY configuration, allowing the MAC to select only the messages it is interested in.

Fig.8 shows a modified version of the 802.16e reference model and the location of the API as described in this document (shown as MAC/PHY API). The 802.16e protocol stack is

separately within a frame configuration structure.

mechanism operates in parallel to this API.

measurements gathered from the PHY or radio subsystem.

**3.4 API architecture design** 

scope of this API document.

Fig. 8. 802.16e Protocol Stack

shown, together with the radio and management entity.

**3.4.1 API introduction** 

separately.

According to the function, the API is divided into two parts: communication with the MAC and the data processing chain of PHY layer. The first part includes getting messages from MAC, sending data to the SDRAM and regrouping the receive message to MAC. The messages from MAC include control-plane messages, which are used to perform configuration and reconfiguration of the PHY, and data-plane messages. Both messages have the same message format which includes message header, which descript message type and PHY entity, and message body. The receive messages to MAC are called response or indication messages. Then it comes to communication with PHY. The input signals of each module in the data processing chain can be classified into two types: control and data signal. Among them, all control signals come from the API; data signal is the output of the previous module. However, the beginning of the data also derives from the API. So, API separately process control and data information to produce the two signals. Their modules are Control system and Data processing, which are shown in Fig. 9.

#### **4. MAC protocol implementation**

#### **4.1 Introduction of MAC functions**

According to the OSI seven-layer network protocol,MAC lays between the PHY layer and the network layer, responsibility for the data convergence and resource scheduling. So there

Design and Implementation of WiMAX Baseband System 255

The main functions of this state machine are to hold the current state of BS, parse the messages which are received from PHY layer, packet the messages which are sent to the PHY layer and transfer to another state. The BS state machine is designed as three states, namely STATE\_CONFIG, STATE\_NORMAL, STATE\_NULL. The BS's initialization state is STATE\_NULL. After the startup, the MAC layer entity of BS will configure the PHY layer with specific parameters and change the current state to STATE\_CONFIG. The state will transfer to STATE\_NORMAL when succeeding the configuration. The state of STATE\_NORMAL indicates that the BS is working normally and any MS can try to access

The main functions of BS-MS-State-Machine are to hold the current state of MSs, parse the messages received from the PHY layer, packet the messages, send to the PHY layer and transfer to other state. According to the possible state of MS, the BS-MS-State-Machine is

After the success of the initial ranging, BS will create a corresponding state machine for the MS. The BS-MS-State-Machine transfers from STATE\_NULL to STATE\_CONNECTED. The MS will execute the process of registration. Once registration succeeds, the state will transfers to STATE\_NORMAL. During the STATE\_NORMAL, MSs can establish the

When the quality of the signals MS received is poor for a long time or the traffic capacity of the BS is saturated, the MS will consider to handover to another BS. Before the handover, the BS-MS-State-Machine will transfer to STATE\_HOSCAN state to scan other BSs. After negotiating with target BSs, the MS will select the best one to process the handover confirm. Once receiving the allowance of the target BS, the BS-MS-State-Machine will transfer to STATE\_HOPROCESS state to execute the handover. The target BS will initial a corresponding BS-MS-State-Machine and transfer to STATE\_HOACCESSED state. The target BS changes into the serving BS. The target BS transfers the state to STATE\_NORMAL and initial serving BS transfer the state to STATE\_HOCOMPLETE. At this point, the scan

defined as having nine states, which are depicted in Fig.12 and Fig.13.

connection with BS to transmit the traffic of video, voice and data services.

BS-State-Machine module

Fig. 11. The state of BS-State-Machine

Fig. 12. The state of BS-MS-State-Machine

and handover process is over.

BS-MS-State-Machine module

to it.

should be appropriate interfaces between the PHY and network layers. In order to highlight the implementation of the MAC layer, for the following description in the chapter, the network layer is designed simply. Its main functions and features are shown in Fig.10 (Du, 2010).


Fig. 10. The structure of MAC functions

#### **4.2 Implementation of MAC layer**

The embedded ARM-Linux operation system is chosen as the development environment of the MAC layer. Considering the requirement of the processing time, the multi-threads techniques are designed so that the MAC layer can packet and parse messages in time. Besides that, the algorithms adopted by the system are needed optimizing to achieve the compromise between system performances and the complexity. Because the message process procedures are different at the different BS/MS state, the state machine is designed to track the state of the BS/MS. The implementation details are introduced in the following part.

#### **4.2.1 State machine**

This section will introduce the BS state machine and MS state machine briefly. For BS, the state machine of BS from the startup to normal and the state of the accessed MS to the current BS are designed. For MS, the corresponding state machine is also designed. The conditions of the state transfer are defined.

1. State machine for BS

According to the functions of BS, the implementation of the BS state machine is divided into two modules. The first one is BS-State-Machine which manages the BS own state and state transition. The second one is MS-State-Machine which manages the states of MSs which have already accessed or attempted to access to the current cell. The BS-State-Machine is responsible for tracking the states of accessed MSs.

should be appropriate interfaces between the PHY and network layers. In order to highlight the implementation of the MAC layer, for the following description in the chapter, the network layer is designed simply. Its main functions and features are shown in Fig.10 (Du,

The embedded ARM-Linux operation system is chosen as the development environment of the MAC layer. Considering the requirement of the processing time, the multi-threads techniques are designed so that the MAC layer can packet and parse messages in time. Besides that, the algorithms adopted by the system are needed optimizing to achieve the compromise between system performances and the complexity. Because the message process procedures are different at the different BS/MS state, the state machine is designed to track the state of the

This section will introduce the BS state machine and MS state machine briefly. For BS, the state machine of BS from the startup to normal and the state of the accessed MS to the current BS are designed. For MS, the corresponding state machine is also designed. The

According to the functions of BS, the implementation of the BS state machine is divided into two modules. The first one is BS-State-Machine which manages the BS own state and state transition. The second one is MS-State-Machine which manages the states of MSs which have already accessed or attempted to access to the current cell. The BS-State-Machine is

BS/MS. The implementation details are introduced in the following part.

2010).

Fig. 10. The structure of MAC functions

conditions of the state transfer are defined.

responsible for tracking the states of accessed MSs.

**4.2 Implementation of MAC layer** 

**4.2.1 State machine** 

1. State machine for BS

The main functions of this state machine are to hold the current state of BS, parse the messages which are received from PHY layer, packet the messages which are sent to the PHY layer and transfer to another state. The BS state machine is designed as three states, namely STATE\_CONFIG, STATE\_NORMAL, STATE\_NULL. The BS's initialization state is STATE\_NULL. After the startup, the MAC layer entity of BS will configure the PHY layer with specific parameters and change the current state to STATE\_CONFIG. The state will transfer to STATE\_NORMAL when succeeding the configuration. The state of STATE\_NORMAL indicates that the BS is working normally and any MS can try to access to it.

Fig. 11. The state of BS-State-Machine

BS-MS-State-Machine module

The main functions of BS-MS-State-Machine are to hold the current state of MSs, parse the messages received from the PHY layer, packet the messages, send to the PHY layer and transfer to other state. According to the possible state of MS, the BS-MS-State-Machine is defined as having nine states, which are depicted in Fig.12 and Fig.13.

After the success of the initial ranging, BS will create a corresponding state machine for the MS. The BS-MS-State-Machine transfers from STATE\_NULL to STATE\_CONNECTED. The MS will execute the process of registration. Once registration succeeds, the state will transfers to STATE\_NORMAL. During the STATE\_NORMAL, MSs can establish the connection with BS to transmit the traffic of video, voice and data services.

Fig. 12. The state of BS-MS-State-Machine

When the quality of the signals MS received is poor for a long time or the traffic capacity of the BS is saturated, the MS will consider to handover to another BS. Before the handover, the BS-MS-State-Machine will transfer to STATE\_HOSCAN state to scan other BSs. After negotiating with target BSs, the MS will select the best one to process the handover confirm. Once receiving the allowance of the target BS, the BS-MS-State-Machine will transfer to STATE\_HOPROCESS state to execute the handover. The target BS will initial a corresponding BS-MS-State-Machine and transfer to STATE\_HOACCESSED state. The target BS changes into the serving BS. The target BS transfers the state to STATE\_NORMAL and initial serving BS transfer the state to STATE\_HOCOMPLETE. At this point, the scan and handover process is over.

Design and Implementation of WiMAX Baseband System 257

When the quality of the MS received signals is poor for a long time or the traffic capacity of the BS is saturated, the MS will consider to handover and send the scan request to serving BS. After receiving the scan response, the MS state will transfer to STATE\_Scanning\_timerM. The MS will set the PHY configuration parameters at STATE\_Scanning\_PHYSyn according to the target BSs, synchronize with the target BS and send the scan result to the serving BS. When all the targets have been scanned, the state

Serving BS will consider the factors of signals strength and so on to select the best target BS to process the handover. The MS state will transfer to STATE\_HO\_Request. The MS will process the downlink synchronization with the target BS at the state of STATE\_HO\_PHY\_SYN and execute the uplink synchronization by ranging mechanism at the state of STATE\_HO\_RANGING. Once having received the successful ranging response, the process of handover finishes. The MS sets the best target BS as the serving BS and

Three threads for MAC protocol are designed, which are time-thread, MAC-thread and APP- thread to keep the system running. The functions performed by the three threads are

Time-thread is mainly used for timing, polling the DMA channels to get the interruption to trigger the next operation. If the interruption type represents data arrival, the MAC layer will get the data from the PHY layer through the API for the further processing. If the interruption type represents frame beginning, then the frame number will increase by 1and the MAC entity will send the packet messages to the PHY entity through the API. Such a send-receive method is aimed to accommodate the requirements of limited time processing. The details can be referred to section 4.3. Because there are many timers in the MAC layer module, the time-thread will also be responsible for the timing and overtime

slightly different at BS and MS sides. The following will make a brief introduction.

transfers to STATE\_MOB\_SCN\_REP.

Fig. 15. MS scanning flow chart

transfers the state to STATE\_NORMAL.

Fig. 16. MS handover flow chart

1. three threads at BS side

**4.2.2 Multi-threading** 

processing.

Fig. 13. The handover state of BS-MS-State-Machine

2. the design of MS state machine

The design principles of MS state machine are the same with those of BS. Similarly, the main functions of MS state machine are saving the current state, parsing the messages receiving from PHY layer, sending the packet messages to PHY layer and transferring the state.

MS-State-Machine module

According to the behaviors of MS, the MS-State-Machine includes several states depicted as Fig.14. During the initial process of MS state machine, the state transfers from STATE\_NULL to STATE\_IDLE. When the MS receives the accessing indication from the upper layer, the MAC entity will process the PHY configuration and transfer the state to STATE\_CONFIG. In order to synchronize with BS, the MS will scan the downlink channels to get downlink synchronization with the BS at the state of STATE\_DL\_SYN and achieve the uplink channel transmitting parameters to send the initial ranging request at the state of STATE\_UL\_SYN. After the synchronization, MS state machine transfers to STATE\_RANGING and executes the process of registration at STATE\_REGISTRATION. The state transfers to STATE\_NORMAL after receiving the successful registration response from BS. The details are shown in Fig.14.

Fig. 14. The normal state setup of MSStaeMachine

When the quality of the MS received signals is poor for a long time or the traffic capacity of the BS is saturated, the MS will consider to handover and send the scan request to serving BS. After receiving the scan response, the MS state will transfer to STATE\_Scanning\_timerM. The MS will set the PHY configuration parameters at STATE\_Scanning\_PHYSyn according to the target BSs, synchronize with the target BS and send the scan result to the serving BS. When all the targets have been scanned, the state transfers to STATE\_MOB\_SCN\_REP.

Fig. 15. MS scanning flow chart

256 Advanced Transmission Techniques in WiMAX

The design principles of MS state machine are the same with those of BS. Similarly, the main functions of MS state machine are saving the current state, parsing the messages receiving

According to the behaviors of MS, the MS-State-Machine includes several states depicted as Fig.14. During the initial process of MS state machine, the state transfers from STATE\_NULL to STATE\_IDLE. When the MS receives the accessing indication from the upper layer, the MAC entity will process the PHY configuration and transfer the state to STATE\_CONFIG. In order to synchronize with BS, the MS will scan the downlink channels to get downlink synchronization with the BS at the state of STATE\_DL\_SYN and achieve the uplink channel transmitting parameters to send the initial ranging request at the state of STATE\_UL\_SYN. After the synchronization, MS state machine transfers to STATE\_RANGING and executes the process of registration at STATE\_REGISTRATION. The state transfers to STATE\_NORMAL after receiving the successful registration response from

from PHY layer, sending the packet messages to PHY layer and transferring the state.

Fig. 13. The handover state of BS-MS-State-Machine

2. the design of MS state machine

MS-State-Machine module

BS. The details are shown in Fig.14.

Fig. 14. The normal state setup of MSStaeMachine

Serving BS will consider the factors of signals strength and so on to select the best target BS to process the handover. The MS state will transfer to STATE\_HO\_Request. The MS will process the downlink synchronization with the target BS at the state of STATE\_HO\_PHY\_SYN and execute the uplink synchronization by ranging mechanism at the state of STATE\_HO\_RANGING. Once having received the successful ranging response, the process of handover finishes. The MS sets the best target BS as the serving BS and transfers the state to STATE\_NORMAL.

Fig. 16. MS handover flow chart

#### **4.2.2 Multi-threading**

Three threads for MAC protocol are designed, which are time-thread, MAC-thread and APP- thread to keep the system running. The functions performed by the three threads are slightly different at BS and MS sides. The following will make a brief introduction.

1. three threads at BS side

Time-thread is mainly used for timing, polling the DMA channels to get the interruption to trigger the next operation. If the interruption type represents data arrival, the MAC layer will get the data from the PHY layer through the API for the further processing. If the interruption type represents frame beginning, then the frame number will increase by 1and the MAC entity will send the packet messages to the PHY entity through the API. Such a send-receive method is aimed to accommodate the requirements of limited time processing. The details can be referred to section 4.3. Because there are many timers in the MAC layer module, the time-thread will also be responsible for the timing and overtime processing.

Design and Implementation of WiMAX Baseband System 259

The interaction between MAC layer and PHY layer is through API entity. The MAC layer parses the API messages received from the API entity and sends the packet API messages to the API entity. The API messages of txstart.request and rxstart.request at the BS side consist of DL-MAP and UL-MAP which are sent in the downlink subframe. The common place of DL-MAP and UL-MAP is that they all contain the frame number. If the frame number is the same, the DL-MAP indicates the current downlink subframe resource allocation and the UL-MAP indicates the current uplink subframe resource allocation. This is referred as minimum-time relevance. If the frame number in the UL-MAP is larger than that in the DL-MAP by1, the DL-MAP indicates the current downlink subframe resource allocation but the UL-MAP indicates the next uplink subframe resource allocation. This is referred as

**4.3 Synchronization between MAC and PHY** 

maximum time relevance. (Zeng, 2006).

Fig. 19. Minimum time relevance of DL-MAP and UL-MAP

Fig. 20. Maximum time relevance of DL-MAP and UL-MAP

the next section.

structure.

**4.3.1 Downlink transmission** 

The API messages should be sent to the API entity before the transmission of the air interface. The processes of downlink and uplink transmission will be briefly introduced in

1. The BS MAC issues a TXSTART.request which includes the TXVECTOR describing the subframe structure according to the schedule. For a valid request, TXSTART.response returns the frame number(N) which it received from the MAC in the TXVECTOR

The construction and transmission of a downlink subframe at BS occurs as follows:

MAC-thread is mainly responsible for processing the messages according to the current state, sending the API messages to PHY layer or getting the API messages from PHY layer.

APP-thread mainly executes the tasks of converging the uplink traffic data and sending the downlink traffic to the MAC-thread which will deal with PDUs forming. The triggers and relations of the three threads are shown in Fig.17

Fig. 17. The triggers and relations of the three threads at BS

2. three threads at MS side

The functions of the three threads at the MS side are similar with BS's. Once the frame interruption arrives, the time-thread will execute the overtime process. And then, the MACthread will process the API messages received from PHY layer according to current state. The MAC-thread will also packet the management messages and the uplink traffic data into PDUs that are sent to PHY layer. The APP-thread is mainly used to converge the downlink traffic data and send the uplink traffic data to the MAC-thread. The triggers and relations of the three threads are shown in Fig.18.

Fig. 18. The triggers and relations of the three threads at MS

#### **4.3 Synchronization between MAC and PHY**

258 Advanced Transmission Techniques in WiMAX

MAC-thread is mainly responsible for processing the messages according to the current state, sending the API messages to PHY layer or getting the API messages from PHY

APP-thread mainly executes the tasks of converging the uplink traffic data and sending the downlink traffic to the MAC-thread which will deal with PDUs forming. The triggers and

The functions of the three threads at the MS side are similar with BS's. Once the frame interruption arrives, the time-thread will execute the overtime process. And then, the MACthread will process the API messages received from PHY layer according to current state. The MAC-thread will also packet the management messages and the uplink traffic data into PDUs that are sent to PHY layer. The APP-thread is mainly used to converge the downlink traffic data and send the uplink traffic data to the MAC-thread. The triggers and relations of

relations of the three threads are shown in Fig.17

Fig. 17. The triggers and relations of the three threads at BS

Fig. 18. The triggers and relations of the three threads at MS

2. three threads at MS side

the three threads are shown in Fig.18.

layer.

The interaction between MAC layer and PHY layer is through API entity. The MAC layer parses the API messages received from the API entity and sends the packet API messages to the API entity. The API messages of txstart.request and rxstart.request at the BS side consist of DL-MAP and UL-MAP which are sent in the downlink subframe. The common place of DL-MAP and UL-MAP is that they all contain the frame number. If the frame number is the same, the DL-MAP indicates the current downlink subframe resource allocation and the UL-MAP indicates the current uplink subframe resource allocation. This is referred as minimum-time relevance. If the frame number in the UL-MAP is larger than that in the DL-MAP by1, the DL-MAP indicates the current downlink subframe resource allocation but the UL-MAP indicates the next uplink subframe resource allocation. This is referred as maximum time relevance. (Zeng, 2006).

Fig. 19. Minimum time relevance of DL-MAP and UL-MAP

Fig. 20. Maximum time relevance of DL-MAP and UL-MAP

The API messages should be sent to the API entity before the transmission of the air interface. The processes of downlink and uplink transmission will be briefly introduced in the next section.

#### **4.3.1 Downlink transmission**

The construction and transmission of a downlink subframe at BS occurs as follows:

1. The BS MAC issues a TXSTART.request which includes the TXVECTOR describing the subframe structure according to the schedule. For a valid request, TXSTART.response returns the frame number(N) which it received from the MAC in the TXVECTOR structure.

Design and Implementation of WiMAX Baseband System 261

3. A MACPDU.response is returned for each request indicating whether the PDU has been successfully queued for transmission. The TXEDN.indication message indicates whether the downlink subframe has been successfully transmitted or if transmission was aborted due to an error. If a TXSTART.request is not received then nothing is transmitted. Therefore to ensure continuous transmission, a TXSTART.request must be

1. The RXSTRAT.request message instructs the PHY to start reception of subframe(N). To allow the PHY to parse the DL-MAP the MAC must send a RXSTART.request message which includes the CIDs to decode. When reception of subframe(N) begins the

2. When the MS is registered with the BS and not in idle or sleep mode, it should issue a

3. MAC PDUs received by the BS are transferred via the MACPDU.indication message. Each MACPDU.indication may contain multiple MACPDUs. Each received PDU is

4. The first MACPDU.indicatiion for burst#0 will contain the FCH, the second for burst#1 will contain the DL-MAP. However, the PHY will parse the FCH and DL-MAP so it is not necessary for the MAC to send any downlink frame structure

5. The end of decoding for each downlink subframe is signaled via the RXEND.indication message. The downlink bursts for the MS may occur early in a subframe so RXEND.indication can be issued before the end of the downlink subframe (PicoChip

1. The MS MAC issues a TXSTART.request which includes the TXVECTOR describing the uplink subframe structure according to schedule. For a valid request, TXSTART. response returns the frame number (N) which it received from the MAC in the

2. The MS MAC issues one MACPUD.request for each uplink PHY burst described in TXVECTOR, there should be only one normal uplink burst allocated to the MS per frame. Each MACPDU.request may contain multiple actual MAC PDUs. The MACPDU.request message must be issued before transmission of the subframe is indicated by TXSTART.indicaation.An error is returned in MACPDU.response if a PDU

3. A MACPDU.response is returned for each request indication whether the PDUs have been successfully queued for transmission, The TXEDN.indication message indicates whether the uplink subframe has been successfully transmitted or if transmission was aborted due to an error. If a TXSTRAT.request is not received then nothing is transmitted. The MAC should only issue the TXSTART.request when it needs to transmit uplink data, perform ranging or send information on the ACK channel or fast-

The construction and transmission of a uplink subframe at MS occurs as follows:

submitted for each and every downlink subframe (PicoChip Company, 2007).

The reception of a downlink subframe at MS occurs as follows:

RXSTRAT.indication message is send to MAC.

associated with the downlink frame number.

feedback channel (PicoChip Company, 2007).

RXSTRAT.request for every subframe.

information to PHY.

Company, 2007).

**4.3.2 Uownlink transmission** 

TXVECTOR structure.

is submitted late.

2. The BS MAC issues one MACPDU.request for each downlink PHY burst(i.e. each burst described in TXVECTOR). It is not necessary for the BS MAC to wait for TXSTART.response before issuing MACPDU.request. Each MACPDU.request may contain multiple actual MAC PDUs. The MACPDU.request messages must be issued in the burst order specified by the TXVECTOR provided with the TXSTART.request and issued before transmission of the subframe is indicated by TXSTART.indication. An error is returned in MACPDU.response if a PDU is submitted too late.

Fig. 21. API primitives for downlink transmission

2. The BS MAC issues one MACPDU.request for each downlink PHY burst(i.e. each burst described in TXVECTOR). It is not necessary for the BS MAC to wait for TXSTART.response before issuing MACPDU.request. Each MACPDU.request may contain multiple actual MAC PDUs. The MACPDU.request messages must be issued in the burst order specified by the TXVECTOR provided with the TXSTART.request and issued before transmission of the subframe is indicated by TXSTART.indication. An

error is returned in MACPDU.response if a PDU is submitted too late.

Fig. 21. API primitives for downlink transmission

3. A MACPDU.response is returned for each request indicating whether the PDU has been successfully queued for transmission. The TXEDN.indication message indicates whether the downlink subframe has been successfully transmitted or if transmission was aborted due to an error. If a TXSTART.request is not received then nothing is transmitted. Therefore to ensure continuous transmission, a TXSTART.request must be submitted for each and every downlink subframe (PicoChip Company, 2007).

The reception of a downlink subframe at MS occurs as follows:


#### **4.3.2 Uownlink transmission**

The construction and transmission of a uplink subframe at MS occurs as follows:


Design and Implementation of WiMAX Baseband System 263

In this chapter, the design and implementation of a WiMAX wireless baseband communication system are presented. Based on the discussion of the typical baseband hardware schemes, we adopt the Picochip multi-core DSP processor as the base of the baseband platform. The hardware, PHY protocol and MAC protocol are introduced in terms of design and implementation other than research aspect, the tradeoff between complexity and performances has been taken into account to meet the requirements. The PHY-MAC interface and API, link-level simulation and debugging method are also mentioned in this chapter, which may provide users better understanding of the

Du, Y. (2010), The MAC layer, In: *IEEE 802.16m Broadband Wireless Technology and System Design*, pp. (181-220), Posts & Telecom Press, Peking, 978-7-115-22754-6. Dohler, M. Lerau, C. &Hardouin, E. (2005). *UMTS FDD multi-Antenna Receiver Complexity* 

Fathi, L. (2004). *Complexity of MPIC receiver for HSDPA R5.* France Telecom R&D, internal

Foschini. G.J.(1996).*Layered Space-Time Architecture for Wireless Communicationin a Fading* 

Goldfarb, M. Palmer, W. & Murphy,T. (2000).*Analog baseband IC for use in direct conversion W-*

Jan Mietzner, Robert Schober, Senior Member, Lutz Lampe, Wolfgang H. Gerstacker and

Jusslia, J. Ryynanen, J. &Kivekas,K. (2001).A 22-ma 3.0-db NF direct conversion receiver for 3g WCDMA.IEEE journal of solid-state Circuits, vol. 36, pp. 2025-2029. LAN/MAN Standards Committee of the IEEE Computer Society and the IEEE Microwave

Parssinen, A. Jussila, J. &Ryynanen, J. (1999).*A 2-GHZ wide-band direct conversion receiver for WCDMA applications.*IEEE Journal of Solid-state Circuits, vol. 34, pp. 1893-1903. Parssinen, A.Jussila, J.&Ryynanen, J. (2000).*A wide-band direct conversion receiver with on-chip* 

S. Lin and P. Yu. (July 1982). *A Hybrid ARQ Scheme with Parity Retransmission for Error Control of Satellite Channel , IEEE Trans on Communications*. pp. 1701–1719.

*CDMA receivers.*Radio Frequency Integrated Circuits (RFIC) Symposium, 2000.

Peter A. Hoeher.(2009). *IEEE COMMUNICATIONS SURVEYS & TUTORIALS,* 

Theory and Techniques Society.(December 2007). *Part 16: Air interface for Broadband* 

*A/D converters.*VLSI Circuits, 2000. Digest of Technical Papers. 2000 Symposium,

.

Bell Labs Technical Journal,

*Estimation.* France Telecom R&D, internal report. pp. 158-175.

*Environment When Using Multi-Element Antennas*

*VOL. 11, NO. 2, SECOND QUARTER 2009, pp*. 87-105.

*Wireless Access System*, pp. (619-1082), P802.16Rev2/D2.

PicoChip Company. (July27,2007). BS MAC-PHY API Definition. pp. 11-13 PicoChip Company. (May21,2007). MS MAC-PHY API Definition. pp. 16-17 PicoChip Company. (September 17,2008).*Tools\_Userdoc\_Fullman\_7.4.5.* pp. 123-138

Digest of Papers. 2000 IEEE, pp. 79-82.

**5. Conclusions** 

development procedure.

report. pp. 204-210.

pp. 32-33.

1996, V01.1(2).pp.41-59.

**6. References** 

Fig. 22. API primitives for uplink transmission

The construction and transmission of an uplink subframe at BS occurs as follows:


#### **5. Conclusions**

262 Advanced Transmission Techniques in WiMAX

Fig. 22. API primitives for uplink transmission

RXEND.indication(PicoChip Company, 2007).

N.

subframe.

The construction and transmission of an uplink subframe at BS occurs as follows:

1. The RXSTART.request message provides the RXVECTOR which describes the bursts on the uplink for frame N. The UL-MAP is transmitted at the start of downlink subframe

2. The RXSTRAT.request is issued against a particular downlink subframe and must be sent after the associated TXSTART.requset. A successful RXSTRAT.request returns the frame number N which is received from the MAC in the RXVECTOR structure. The

4. MAC PDUs received by the BS are transferred via the MACPDU.indication message. Each MACPDU.indication may contain multiple MAC PDUs. Each received PDU is associated with uplink frame number and the burst number specified in RXVECTOR. The end of decoding for each uplink subframe is signaled via the

RXSTRAT.request must also be issued before the start of downlink subframe N. 3. If a RXSTART.request is not received then the receiver is effectively disabled. Therefore to ensure continuous reception, a RXSTART.request must be submitted for each uplink In this chapter, the design and implementation of a WiMAX wireless baseband communication system are presented. Based on the discussion of the typical baseband hardware schemes, we adopt the Picochip multi-core DSP processor as the base of the baseband platform. The hardware, PHY protocol and MAC protocol are introduced in terms of design and implementation other than research aspect, the tradeoff between complexity and performances has been taken into account to meet the requirements. The PHY-MAC interface and API, link-level simulation and debugging method are also mentioned in this chapter, which may provide users better understanding of the development procedure.

#### **6. References**


**13** 

*Egypt* 

**Performance Evaluation of WiMAX System** 

In this chapter, we introduce a new class of coding technique that belongs to product code family. This technique is based on convolutional code. The use of convolutional code in the product code setting makes it possible to use the vast knowledge base for convolutional

Product codes studied thus far have been constructed using linear block codes, such as Hamming [1], Bose–Chaudhuri–Hocquenghem (BCH) [2] and [3], Reed Solomon codes [4] and single parity check (SPC) [5]. These types of the product codes are traditionally constructed by linear block codes that have structure with a time varying property [6].

The product code proposed in this chapter is constructed by using time-invariant convolutional code. Its component codes' trellis structure does not vary in time as in product codes constructed with Hamming, BCH, and Reed Solomon block codes. Moreover, the number of states in the trellis structure of a block code may grow exponentially with the difference of codeword and data block lengths, whereas the number of states in a

The time invariant trellis structure of convolutional codes makes them more convenient for implementation. In addition, numerous practical techniques such as trellis coded modulation and puncturing can be simply utilized with convolutional codes as opposed to

Multi-input multi-output (MIMO) techniques are quite important to enhance the capacity of wireless communication systems. Space-time trellis codes provide both diversity and coding gain in MIMO channels and are widely used [7]. Space-time trellis codes usually have timeinvariant trellis structures just like convolutional codes. Thus, a product code based on convolutional codes is more suitable for integration with MIMO channels and poses an

The type of proposed product code described in this chapter is called modified Convolutional Product Codes (CPC), considered as a different type of normal CPC [8]. The normal CPC depends on recursive systematic convolutional encoder, whereas the modified version of CPC will basically depend on non-recursive non-systematic convolutional

**1. Introduction** 

linear block codes.

encoder.

codes as well as their flexibility.

convolutional code can be set as desired.

alternative to block product codes.

**Using Different Coding Techniques** 

M. Shokair, A. Ebian, and K. H. Awadalla

*El-Menoufia University,* 

Zeng, C. (2006), The support from MAC layer to PHY layer, In: The Principles and Applications of WIMAX/802.16, pp. (113-115), China Machine Press, Peking, 7- 11120111-6.

### **Performance Evaluation of WiMAX System Using Different Coding Techniques**

M. Shokair, A. Ebian, and K. H. Awadalla *El-Menoufia University, Egypt* 

#### **1. Introduction**

264 Advanced Transmission Techniques in WiMAX

Zeng, C. (2006), The support from MAC layer to PHY layer, In: The Principles and

11120111-6.

Applications of WIMAX/802.16, pp. (113-115), China Machine Press, Peking, 7-

In this chapter, we introduce a new class of coding technique that belongs to product code family. This technique is based on convolutional code. The use of convolutional code in the product code setting makes it possible to use the vast knowledge base for convolutional codes as well as their flexibility.

Product codes studied thus far have been constructed using linear block codes, such as Hamming [1], Bose–Chaudhuri–Hocquenghem (BCH) [2] and [3], Reed Solomon codes [4] and single parity check (SPC) [5]. These types of the product codes are traditionally constructed by linear block codes that have structure with a time varying property [6].

The product code proposed in this chapter is constructed by using time-invariant convolutional code. Its component codes' trellis structure does not vary in time as in product codes constructed with Hamming, BCH, and Reed Solomon block codes. Moreover, the number of states in the trellis structure of a block code may grow exponentially with the difference of codeword and data block lengths, whereas the number of states in a convolutional code can be set as desired.

The time invariant trellis structure of convolutional codes makes them more convenient for implementation. In addition, numerous practical techniques such as trellis coded modulation and puncturing can be simply utilized with convolutional codes as opposed to linear block codes.

Multi-input multi-output (MIMO) techniques are quite important to enhance the capacity of wireless communication systems. Space-time trellis codes provide both diversity and coding gain in MIMO channels and are widely used [7]. Space-time trellis codes usually have timeinvariant trellis structures just like convolutional codes. Thus, a product code based on convolutional codes is more suitable for integration with MIMO channels and poses an alternative to block product codes.

The type of proposed product code described in this chapter is called modified Convolutional Product Codes (CPC), considered as a different type of normal CPC [8]. The normal CPC depends on recursive systematic convolutional encoder, whereas the modified version of CPC will basically depend on non-recursive non-systematic convolutional encoder.

Performance Evaluation of WiMAX System Using Different Coding Techniques 267

The general encoding procedure, which includes any type of convolutional encoder and

interleaver, is illustrated in Figures 3 & 4.

Fig. 3. CPC Encoding Procedure without an Interleaver.

Fig. 4. Convolutional Product Code Encoder with any type of

In the decoding process, the log-MAP soft decoding algorithm, [9] and [10], is used to iteratively decode the convolutional product code. Since columns were encoded last, each column is independently decoded one by one. The extrinsic information obtained from the columns is passed to the row decoder after being de-interleaved. Then, row decoding proceeds; rows are decoded one by one, and interleaved extrinsic information is passed to

The decoding structure employed in this method is the same as that of serially-concatenated

In the modified version of CPC, the same technique is used for coding the message, except using nonrecursive nonsystematic convolutional encoder instead of recursive systematic

the column decoder. The CPC decoding procedure is depicted in Figure 5.

Interleaver (d denotes data bits and p denotes parity bits).

**3. CPC decoder** 

codes in Figure 6.6 [11].

**4. Modified CPC encoder** 

WiMAX system is a wireless communication system. It suffers from having a high *Bit Error Rate* (BER) at low *Signal to Noise Ratio* (SNR). Using modified version of CPC for WiMAX system reduces the BER at low SNR. Also using the modified version of CPC with WiMAX decreases the number of stages of its physical layer as described later.

#### **2. CPC encoder**

For a regular product code, the information bits are placed into a matrix. The rows and columns are encoded separately using linear block codes. This type of a product encoder is shown in Figure 1.

Fig. 1. Regular Product Code Encoding Procedure.

In CPC method, the information bits are placed into two dimensions (2D) matrix. The rows and the columns are encoded separately using recursive systematic convolutional encoders. Each row of the matrix is encoded using a convolutional code with generator polynomial (1, 5/7) octal and code rate (1/2) Figure 2. The same recursive systematic convolutional code with the same polynomial is used to encode each row. Once all rows have been encoded, the matrix is sent, if desired, to an interleaver. The original data matrix dimensions are (n × k), and the encoded data matrix dimensions will be (2n×k) for coding rate (1/2).

The coded rows matrix is then recoded column by column using the same or different recursive systematic convolutional encoder. Hence, the overall code rate is 1/4.

Fig. 2. CPC Convolutional Coding [1 , 5/7].

The general encoding procedure, which includes any type of convolutional encoder and interleaver, is illustrated in Figures 3 & 4.

Fig. 3. CPC Encoding Procedure without an Interleaver.

Fig. 4. Convolutional Product Code Encoder with any type of

Interleaver (d denotes data bits and p denotes parity bits).

#### **3. CPC decoder**

266 Advanced Transmission Techniques in WiMAX

WiMAX system is a wireless communication system. It suffers from having a high *Bit Error Rate* (BER) at low *Signal to Noise Ratio* (SNR). Using modified version of CPC for WiMAX system reduces the BER at low SNR. Also using the modified version of CPC with WiMAX

For a regular product code, the information bits are placed into a matrix. The rows and columns are encoded separately using linear block codes. This type of a product encoder is

In CPC method, the information bits are placed into two dimensions (2D) matrix. The rows and the columns are encoded separately using recursive systematic convolutional encoders. Each row of the matrix is encoded using a convolutional code with generator polynomial (1, 5/7) octal and code rate (1/2) Figure 2. The same recursive systematic convolutional code with the same polynomial is used to encode each row. Once all rows have been encoded, the matrix is sent, if desired, to an interleaver. The original data matrix dimensions are (n × k),

The coded rows matrix is then recoded column by column using the same or different

and the encoded data matrix dimensions will be (2n×k) for coding rate (1/2).

recursive systematic convolutional encoder. Hence, the overall code rate is 1/4.

decreases the number of stages of its physical layer as described later.

Fig. 1. Regular Product Code Encoding Procedure.

Fig. 2. CPC Convolutional Coding [1 , 5/7].

**2. CPC encoder** 

shown in Figure 1.

In the decoding process, the log-MAP soft decoding algorithm, [9] and [10], is used to iteratively decode the convolutional product code. Since columns were encoded last, each column is independently decoded one by one. The extrinsic information obtained from the columns is passed to the row decoder after being de-interleaved. Then, row decoding proceeds; rows are decoded one by one, and interleaved extrinsic information is passed to the column decoder. The CPC decoding procedure is depicted in Figure 5.

The decoding structure employed in this method is the same as that of serially-concatenated codes in Figure 6.6 [11].

#### **4. Modified CPC encoder**

In the modified version of CPC, the same technique is used for coding the message, except using nonrecursive nonsystematic convolutional encoder instead of recursive systematic

Performance Evaluation of WiMAX System Using Different Coding Techniques 269

2. Then each row of the resulting matrix will be coded by the same generator polynomials. From Figure 7 since the generator polynomials used for coding both rows and columns are (5,7)octal with constraint length 3, not following the standard of WiMAX, each column is padded with two zeros for terminating its encoder. But each row is padded with two or three zeros according to the number of used subcarriers, 128 or 512, receptively to form the suitable size of the overall matrix. That matrix is then divided into smaller matrices with

After the coding process, the total number of bits will be more than the original message bits due to the increase in the overall code rate (1/4), and the addition of the zeros in both column and rows that used for the termination process. Therefore the following steps are

1. Dividing the overall matrix produced from modified CPC into three matrices. Each one has a size (nx4) or (nx6) according to the type of QAM used as mentioned before. The reason for using three matrices only is to have a number of message bits equals to bits used in the convolutional code method, as a comparison between it and CPC is

4. Performing the IFFT on the three matrixes independently resulting in three OFDMA

The reason for using nonrecursive nonsystematic convolutional encoder instead of recursive systematic convolutional encoders is simplifying the termination of the encoder, as RSC contains a feedback and its termination will be more difficult. Also using the generator polynomials (5,7) leads to a little increase in the complexity of the system because of a few

2. Applying symbol mapping for each one independently (16QAM or 64 QAM).

3. Inserting the pilot and DC subcarriers for each matrix.

number of zeros will be added to terminate the two encoders.

5. Applying (cyclic prefix) CP for each symbol. 6. Sending each symbol independently.

The coding by modified CPC will be done in 2 stages

1. Each column will be independently coded.

sizes (nx4) or (nx6) as described later.

Fig. 7. Convolutional Coding [5,7].

done,

done.

symbols.

convolutional encoders for coding both rows and columns. That means the both encoders of rows and columns will have coding rate (1/2), and generator polynomial (5,7) Octa Figure 7.

Fig. 5. Decoding Operation of the Convolutional Product Code.

Fig. 6. Serial Encoding & Decoding Operations

The sequence of bits is fed into 2D matrix and fills it column by column. The size of this matrix depends only on the type of modulation used. For 16 QAM, the size of the matrix will be (nx4) and for 64 QAM the size of the matrix will be (nx6). These sizes simplify the process of mapping, as the symbol size in 16 QAM is 4 bits and in 64 QAM is 6 bits. So each row of those matrices will form one QAM symbol. The 'n' refers to the number of data subcarriers of OFDMA, 128 or 512.

The coding by modified CPC will be done in 2 stages

1. Each column will be independently coded.

268 Advanced Transmission Techniques in WiMAX

convolutional encoders for coding both rows and columns. That means the both encoders of rows and columns will have coding rate (1/2), and generator polynomial (5,7) Octa Figure 7.

The sequence of bits is fed into 2D matrix and fills it column by column. The size of this matrix depends only on the type of modulation used. For 16 QAM, the size of the matrix will be (nx4) and for 64 QAM the size of the matrix will be (nx6). These sizes simplify the process of mapping, as the symbol size in 16 QAM is 4 bits and in 64 QAM is 6 bits. So each row of those matrices will form one QAM symbol. The 'n' refers to the number of data

Fig. 5. Decoding Operation of the Convolutional Product Code.

Fig. 6. Serial Encoding & Decoding Operations

subcarriers of OFDMA, 128 or 512.

2. Then each row of the resulting matrix will be coded by the same generator polynomials.

From Figure 7 since the generator polynomials used for coding both rows and columns are (5,7)octal with constraint length 3, not following the standard of WiMAX, each column is padded with two zeros for terminating its encoder. But each row is padded with two or three zeros according to the number of used subcarriers, 128 or 512, receptively to form the suitable size of the overall matrix. That matrix is then divided into smaller matrices with sizes (nx4) or (nx6) as described later.

Fig. 7. Convolutional Coding [5,7].

After the coding process, the total number of bits will be more than the original message bits due to the increase in the overall code rate (1/4), and the addition of the zeros in both column and rows that used for the termination process. Therefore the following steps are done,


The reason for using nonrecursive nonsystematic convolutional encoder instead of recursive systematic convolutional encoders is simplifying the termination of the encoder, as RSC contains a feedback and its termination will be more difficult. Also using the generator polynomials (5,7) leads to a little increase in the complexity of the system because of a few number of zeros will be added to terminate the two encoders.

Performance Evaluation of WiMAX System Using Different Coding Techniques 271

6. Space-time trellis codes usually have time-invariant trellis structures just like convolutional codes. Thus, a product code based on convolutional codes is more

But on the other hand it causes more delay for obtaining the original message because the code rate becomes 1/4 not 1/2 as in convolutional code. The performance of the system will

This section contains comparisons between the modified CPC method and convolutional code, turbo code and LDPC code. Several results obtained at different types of the channels, modulation techniques (16QAM – 64QAM) and number of OFDM subcarriers (128 -512).

In this work, a matlab tool is used to simulate the physical layer of WiMAX and apply the

In this section the coded signal is transmitted through AWGN channel only. This can be done using matlab function **AWGN.** The syntax of this function is: *y = awgn (x, snr , 'measured')* adds white Gaussian noise to the vector signal **x** to produce output signal **y** The scalar snr specifies the signal-to-noise ratio per sample, in dB. If x is complex, awgn adds

We can derive the relationship between Es/N0 and SNR for complex input signals as

A good rule of thumb for selecting the symbol period value is to set it to be what we model as the symbol period in the model. The value would depend upon what constitutes a symbol and what the oversampling applied to. From Figure 8 to Figure 13 BER versus different received SNR values are shown for the comparison between modified CPC and

(1)

be reduced and this is the price to be paid for the improvement obtained.

complex noise. This syntax measures the power of x before adding noise**.** 

suitable for integration with MIMO channels.

7. Increasing the free distance to be d 2min.

**8. Results** 

mentioned coding methods.

S = Input signal power, in watts

Bn = Noise bandwidth, in Hertz

Tsym = The signal's symbol period.

Fs = Sampling frequency, in Hertz = 1/Tsamp.

Tsamp = The period of each row of a frame-based matrix.

N = Noise power, in watts

**8.1 AWGN channel** 

follows:

Where

#### **5. Modified CPC decoder**

At the receiver, the three OFDMA symbols are combined to form the original matrix which is decoded by Viterbi decoder. The Viterbi decoder uses the same generator polynomials (5,7) with hard decision for each row and for each column. The rows must be decoded first then the columns are done, because columns are encoded first Figure 6. To match the CPC method, the number of data bits will be reduced. For example in OFDMA (128-16QAM) and (128-64QAM) the number of data bits was 144 and 216 but in CPC method it becomes 136 and 204 bits receptively due to the number of zero bits added to terminate the two encoders.

#### **6. Modified CPC minimum distance and its asymptotic performance**

The Hamming weight of a binary codeword is defined as the number of '1's available in the codeword. The minimum distance of a linear code is the minimum Hamming weight of all the codewords. The minimum distance plays an important role in the code performance. As it gets larger, code performance improves, especially at high signal-to-noise ratio (SNR) values. The free distance of the component convolutional codes used in modified CPC with trellis termination will be called dfree. The minimum distance of the modified CPC in the case of no interleavers will be investigated.

#### **No Interleaving**

After the first stage of the modified CPC encoding operation (columns encoding), it is obvious that one of these columns should contain at least dfree number of '1's. This means that there are dfree rows containing at least a *single* '1' in the columns-encoded matrix. When rows are encoded, there exists at least dfree number of rows each containing at least dfree '1's. Hence, in total there are at least d 2min '1's in the coded matrix. In summary, if no interleavers are used, the modified CPC minimum distance is d 2 min.

#### **7. Advantage and disadvantage of CPC**

CPC technique has mainly two main advantages that make it a motivating step for future considerations and improvements for practical systems.


But on the other hand it causes more delay for obtaining the original message because the code rate becomes 1/4 not 1/2 as in convolutional code. The performance of the system will be reduced and this is the price to be paid for the improvement obtained.

#### **8. Results**

270 Advanced Transmission Techniques in WiMAX

At the receiver, the three OFDMA symbols are combined to form the original matrix which is decoded by Viterbi decoder. The Viterbi decoder uses the same generator polynomials (5,7) with hard decision for each row and for each column. The rows must be decoded first then the columns are done, because columns are encoded first Figure 6. To match the CPC method, the number of data bits will be reduced. For example in OFDMA (128-16QAM) and (128-64QAM) the number of data bits was 144 and 216 but in CPC method it becomes 136 and 204 bits receptively due to the number of zero bits added to

The Hamming weight of a binary codeword is defined as the number of '1's available in the codeword. The minimum distance of a linear code is the minimum Hamming weight of all the codewords. The minimum distance plays an important role in the code performance. As it gets larger, code performance improves, especially at high signal-to-noise ratio (SNR) values. The free distance of the component convolutional codes used in modified CPC with trellis termination will be called dfree. The minimum distance of the modified CPC in the

After the first stage of the modified CPC encoding operation (columns encoding), it is obvious that one of these columns should contain at least dfree number of '1's. This means that there are dfree rows containing at least a *single* '1' in the columns-encoded matrix. When rows are encoded, there exists at least dfree number of rows each containing at least dfree '1's.

CPC technique has mainly two main advantages that make it a motivating step for future

1. Do not need another interleaver after channel coding because of converting into matrix (nx4) or (nx6) does almost the same job as the overall matrix will be filled column by column and will be read row by row after coding processes (block interleaver) since

3. The product code we propose in CPC is constructed by using time invariant convolutional codes. Its component codes' trellis structure does not vary in time as in product codes constructed with Hamming, extended Hamming, BCH, and Reed Solomon block codes. The time invariant trellis structure of convolutional codes makes

5. Numerous practical techniques such as trellis coded modulation and puncturing can be

4. The number of states in CPC like a convolutional code can be set as desired.

simply utilized with convolutional codes as opposed to linear block codes.

min '1's in the coded matrix. In summary, if no interleavers

**6. Modified CPC minimum distance and its asymptotic performance** 

**5. Modified CPC decoder** 

terminate the two encoders.

**No Interleaving** 

case of no interleavers will be investigated.

are used, the modified CPC minimum distance is d 2min.

considerations and improvements for practical systems.

each row is used for making QAM symbol.

them more convenient for implementation

**7. Advantage and disadvantage of CPC** 

Hence, in total there are at least d 2

2. Reducing the BER at low SNR.

This section contains comparisons between the modified CPC method and convolutional code, turbo code and LDPC code. Several results obtained at different types of the channels, modulation techniques (16QAM – 64QAM) and number of OFDM subcarriers (128 -512).

In this work, a matlab tool is used to simulate the physical layer of WiMAX and apply the mentioned coding methods.

#### **8.1 AWGN channel**

In this section the coded signal is transmitted through AWGN channel only. This can be done using matlab function **AWGN.** The syntax of this function is: *y = awgn (x, snr , 'measured')* adds white Gaussian noise to the vector signal **x** to produce output signal **y** The scalar snr specifies the signal-to-noise ratio per sample, in dB. If x is complex, awgn adds complex noise. This syntax measures the power of x before adding noise**.** 

We can derive the relationship between Es/N0 and SNR for complex input signals as follows:

$$E\_s \, / N\_0 \, \text{(dB)} = 10 \log\_{10} \left( (S \cdot T\_{sym}) / (N \, / B\_n) \right)$$

$$= 10 \log\_{10} \left( (T\_{sym} F\_s) \cdot (S \, / N) \right)$$

$$= 10 \log\_{10} \left( T\_{sym} \, / T\_{samp} \right) + \text{SNR} \, (\text{dB}) \tag{1}$$

Where

S = Input signal power, in watts

N = Noise power, in watts

Bn = Noise bandwidth, in Hertz

Fs = Sampling frequency, in Hertz = 1/Tsamp.

Tsamp = The period of each row of a frame-based matrix.

Tsym = The signal's symbol period.

A good rule of thumb for selecting the symbol period value is to set it to be what we model as the symbol period in the model. The value would depend upon what constitutes a symbol and what the oversampling applied to. From Figure 8 to Figure 13 BER versus different received SNR values are shown for the comparison between modified CPC and

Performance Evaluation of WiMAX System Using Different Coding Techniques 273

From Figure 10 and Figure 11, the results of comparisons between modified CPC and LDPC code are shown at different received SNR values. From these figures, we conclude that modified CPC gives good results at different SNR. Figure 10 shows that modified CPC coding technique gives better results than LDPC coding technique at 16 QAM and OFDM subcarriers equals 128. The improvement is more than 3 dB for 10 -3. Also, there will be an improvement, when the number of subcarriers is increased to 512 as shown in

Fig. 10. BER Comparison between LDPC code, CPC at 16 QAM, N=128.

Fig. 11. BER Comparison between LDPC code, CPC at 16 QAM, N=512

Figure 11.

convolutional code, LDPC code and turbo code respectively. These comparisons are obtained for modulation type 16QAM and number of subcarriers equals 128 and 512 respectively.

The comparisons between modified CPC and convolutional code are shown in both Figure 8 and Figure 9. From Figure 8, it is shown that SNR will be improved by approximately 2 dB at BER equals 10 -3 for modulation type 16QAM and number of subcarriers equals 128. Also, an improvement can be obtained when the number of subcarriers is increased to 512 as shown in Figure 9.

Fig. 8. BER Comparison between Conv code, CPC at 16 QAM, N=128.

Fig. 9. BER Comparison between Conv code, CPC at 16QAM, N=512.

convolutional code, LDPC code and turbo code respectively. These comparisons are obtained for modulation type 16QAM and number of subcarriers equals 128 and 512

The comparisons between modified CPC and convolutional code are shown in both Figure 8 and Figure 9. From Figure 8, it is shown that SNR will be improved by approximately 2 dB at BER equals 10 -3 for modulation type 16QAM and number of subcarriers equals 128. Also, an improvement can be obtained when the number of subcarriers is increased to 512 as

Fig. 8. BER Comparison between Conv code, CPC at 16 QAM, N=128.

Fig. 9. BER Comparison between Conv code, CPC at 16QAM, N=512.

respectively.

shown in Figure 9.

From Figure 10 and Figure 11, the results of comparisons between modified CPC and LDPC code are shown at different received SNR values. From these figures, we conclude that modified CPC gives good results at different SNR. Figure 10 shows that modified CPC coding technique gives better results than LDPC coding technique at 16 QAM and OFDM subcarriers equals 128. The improvement is more than 3 dB for 10 -3. Also, there will be an improvement, when the number of subcarriers is increased to 512 as shown in Figure 11.

Fig. 10. BER Comparison between LDPC code, CPC at 16 QAM, N=128.

Fig. 11. BER Comparison between LDPC code, CPC at 16 QAM, N=512

Performance Evaluation of WiMAX System Using Different Coding Techniques 275

The comparison for modulation type 64 QAM between modified CPC and convolutional code, LDPC code and turbo code are shown through Figure 14 to Figure 19. BER versus different received SNR values are shown in these figures. These comparisons are obtained and number of subcarriers equals 128 and 512 respectively. The comparisons between

modified CPC and convolutional code are shown in both Figure 14 and Figure 15.

Fig. 14. BER Comparison between Conv code, CPC at 64 QAM, N=128.

Fig. 15. BER Comparison between Conv code, CPC at 64 QAM, N=512.

From Figure 14, it is shown that SNR will be improved by approximately 1.5 dB at BER equals 10 -2 for modulation type 64QAM and number of subcarriers equals 128. Also, an

From Figure 12 to Figure 13, the results produced from the comparisons between modified CPC and turbo code are shown at different received SNR values. As shown from these figures, modified CPC method gives good results compared to turbo coding. Figure 12 shows that using modified CPC method can give better results than turbo coding technique at 16QAM and OFDM subcarriers equals 128. This improvement is more than 3 dB for BER= 10 -3. Also other improvements can be obtained at different number of OFDM subcarriers (512) as shown in Figure 13.

Fig. 12. BER Comparison between Turbo code, CPC at 16 QAM, N=128.

Fig. 13. BER Comparison between Turbo code, CPC 16 QAM, N=512.

From Figure 12 to Figure 13, the results produced from the comparisons between modified CPC and turbo code are shown at different received SNR values. As shown from these figures, modified CPC method gives good results compared to turbo coding. Figure 12 shows that using modified CPC method can give better results than turbo coding technique at 16QAM and OFDM subcarriers equals 128. This improvement is more than 3 dB for BER= 10 -3. Also other improvements can be obtained at different number of OFDM subcarriers

Fig. 12. BER Comparison between Turbo code, CPC at 16 QAM, N=128.

Fig. 13. BER Comparison between Turbo code, CPC 16 QAM, N=512.

(512) as shown in Figure 13.

The comparison for modulation type 64 QAM between modified CPC and convolutional code, LDPC code and turbo code are shown through Figure 14 to Figure 19. BER versus different received SNR values are shown in these figures. These comparisons are obtained and number of subcarriers equals 128 and 512 respectively. The comparisons between modified CPC and convolutional code are shown in both Figure 14 and Figure 15.

Fig. 14. BER Comparison between Conv code, CPC at 64 QAM, N=128.

Fig. 15. BER Comparison between Conv code, CPC at 64 QAM, N=512.

From Figure 14, it is shown that SNR will be improved by approximately 1.5 dB at BER equals 10 -2 for modulation type 64QAM and number of subcarriers equals 128. Also, an

Performance Evaluation of WiMAX System Using Different Coding Techniques 277

From Figure 18 to Figure 19, the results produced from the comparisons between modified CPC and turbo code are shown at different received SNR values. As shown from these figures, modified CPC method gives good results compared to turbo coding. Figure 18 shows that using modified CPC method can give better results than turbo coding technique at 64QAM and OFDM subcarriers equals 128. This improvement is about than 2 dB for 10 -2. Also other improvements can be obtained at different number of OFDM subcarriers (512) as

Fig. 18. BER Comparison between Turbo code, CPC at 64 QAM, N=128.

Fig. 19. BER Comparison between Turbo code, CPC 64 QAM, N=512.

shown in Figure 19.

improvement can be obtained when the number of subcarriers is increased to 512 as shown in Figure 15. Figure 16 and Figure 17 show the results of comparisons between modified CPC and LDPC code are shown at different received SNR values. We conclude that modified CPC gives good results at different SNR. Figure 16 shows that modified CPC coding technique gives better results than LDPC coding technique at 64QAM and OFDM subcarriers equals 128. The improvement is more than 1.5 dB for 10 -2. Also, there will be an improvement, when the number of subcarriers is increased to 512 as shown in Figure 17.

Fig. 16. BER Comparison between LDPC code, CPC at 64QAM, N=128.

Fig. 17. BER Comparison between LDPC code, CPC at 64 QAM, N=512.

improvement can be obtained when the number of subcarriers is increased to 512 as shown in Figure 15. Figure 16 and Figure 17 show the results of comparisons between modified CPC and LDPC code are shown at different received SNR values. We conclude that modified CPC gives good results at different SNR. Figure 16 shows that modified CPC coding technique gives better results than LDPC coding technique at 64QAM and OFDM subcarriers equals 128. The improvement is more than 1.5 dB for 10 -2. Also, there will be an improvement, when the number of subcarriers is increased to 512 as shown in Figure 17.

Fig. 16. BER Comparison between LDPC code, CPC at 64QAM, N=128.

Fig. 17. BER Comparison between LDPC code, CPC at 64 QAM, N=512.

From Figure 18 to Figure 19, the results produced from the comparisons between modified CPC and turbo code are shown at different received SNR values. As shown from these figures, modified CPC method gives good results compared to turbo coding. Figure 18 shows that using modified CPC method can give better results than turbo coding technique at 64QAM and OFDM subcarriers equals 128. This improvement is about than 2 dB for 10 -2. Also other improvements can be obtained at different number of OFDM subcarriers (512) as shown in Figure 19.

Fig. 18. BER Comparison between Turbo code, CPC at 64 QAM, N=128.

Fig. 19. BER Comparison between Turbo code, CPC 64 QAM, N=512.

Performance Evaluation of WiMAX System Using Different Coding Techniques 279

From Figure 20 to Figure 25 BER versus different received SNR values are shown for the comparison between modified CPC and convolutional code, LDPC code and turbo code respectively through the fading channel. These comparisons are obtained for modulation

The comparisons between modified CPC and convolutional code are shown in both Figures 20 and 21. In Figure 20, it is shown that SNR is improved by more than 4 dB at BER equals 10-2 for the number of subcarriers equals 128. An improvement is obtained if the number of

type 16QAM and number of subcarriers equals 128 and 512 respectively.

Fig. 20. BER Comparison between Conv code, CPC at 16QAM , N=128.

Fig. 21. BER Comparison between Conv code, CPC at 16 QAM, N=512.

subcarriers is increased to 512 as shown in Figure 21.

#### **8.2 AWGN plus fading channel**

In this section, the transmitted signal is assumed to pass through time selectivity fading channel plus AWGN. This is done using matlab function **rayleighchan**. The syntax of this function is *chan = rayleighchan (Ts,Fd)* that constructs a frequency-flat ("single path") Rayleigh fading channel object.

Ts is the sample time of the input signal, in seconds. Fd is the maximum Doppler shift, in Hertz.

$$\text{Sample time} = 1/\text{(channel bandwidth x28/25)}\tag{2}$$

$$\text{Maximum Doppler shift} = \mathbf{F}\_{\text{d}} = \begin{pmatrix} v \ \mathbf{f}\_{\text{c}} \end{pmatrix} / \ \mathbf{C}\_{0} \tag{3}$$

Where *fc* is carrier frequency, *v* is the maximum speed between transmitter and the receiver and *C*0 is the speed of light. The Rayleigh multipath fading channel simulators of this toolbox use the band-limited discrete multipath channel model. It is assumed that the delay power profile and the Doppler spectrum of the channel are separable. The multipath fading channel is therefore modeled as a linear finite impulse-response (FIR) filter. Let *Si* denotes the set of samples at the input to the channel. Then the samples *yi* at the output of the channel are related to *Si* through:

$$\mathbf{y}\_{i} = \sum\_{n=-N\_{1}}^{N\_{1}} s\_{i-n} \mathbf{g}\_{n} \tag{4}$$

Where *gn* is the set of tap weights given by:

$$\mathcal{G}\_n = \sum\_{k=1}^K a\_k \text{sinc}\left[\frac{\mathbf{r}\_k}{T\_s} - n\right], \ -N\_1 \le n \le N\_2 \tag{5}$$

In the equations above:


This simulation is done for different coding techniques that have different coding rates because we follow the standard in our simulation. The following parameters are used in our simulation:


In this section, the transmitted signal is assumed to pass through time selectivity fading channel plus AWGN. This is done using matlab function **rayleighchan**. The syntax of this function is *chan = rayleighchan (Ts,Fd)* that constructs a frequency-flat ("single path")

Ts is the sample time of the input signal, in seconds. Fd is the maximum Doppler shift, in

Where *fc* is carrier frequency, *v* is the maximum speed between transmitter and the receiver and *C*0 is the speed of light. The Rayleigh multipath fading channel simulators of this toolbox use the band-limited discrete multipath channel model. It is assumed that the delay power profile and the Doppler spectrum of the channel are separable. The multipath fading channel is therefore modeled as a linear finite impulse-response (FIR) filter. Let *Si* denotes the set of samples at the input to the channel. Then the samples *yi* at the output of the

*τk* ,where 1≤ k ≤ K, is the set of path delays. K is the total number of paths in the

*ak* ,where 1≤ k ≤ K, is the set of complex path gains of the multipath fading channel.

This simulation is done for different coding techniques that have different coding rates because we follow the standard in our simulation. The following parameters are used in our

*N1* and *N2* are chosen so that | *gn* |is small when n is less than *N1* or greater than *N2*

2. Channel Bandwidth (1.25 MHz for IFFT size=128 and 5.00 MHz for IFFT size= 512).

6. Convolutional code with rate equals (1/2), turbo code with rate equals (2/3) and LDPC

Sample time =1/(channel bandwidth x28/25) (2)

Maximum Doppler shift = Fd = (*v* fc) / C0 (3)

(4)

(5)

**8.2 AWGN plus fading channel** 

Rayleigh fading channel object.

channel are related to *Si* through:

In the equations above:

simulation:

multipath fading channel.

1. Frequency band is 3.5 GHz.

4. Oversampling rate is 28/25. 5. Max speed 120 Kmph.

code with rate equals (1/2).

3. Modulation types (16 QAM, 64 QAM).

Where *gn* is the set of tap weights given by:

*Ts* is the input sample period to the channel.

These path gains are uncorrelated with each other.

Hertz.

From Figure 20 to Figure 25 BER versus different received SNR values are shown for the comparison between modified CPC and convolutional code, LDPC code and turbo code respectively through the fading channel. These comparisons are obtained for modulation type 16QAM and number of subcarriers equals 128 and 512 respectively.

The comparisons between modified CPC and convolutional code are shown in both Figures 20 and 21. In Figure 20, it is shown that SNR is improved by more than 4 dB at BER equals 10-2 for the number of subcarriers equals 128. An improvement is obtained if the number of subcarriers is increased to 512 as shown in Figure 21.

Fig. 20. BER Comparison between Conv code, CPC at 16QAM , N=128.

Fig. 21. BER Comparison between Conv code, CPC at 16 QAM, N=512.

Performance Evaluation of WiMAX System Using Different Coding Techniques 281

SNR values for modulation type 16QAM. There is an improvement in SNR by more than 8 dB at BER equals 10-2 for 16QAM and number of OFDMA subcarriers equals 128, this is shown from Figure 24. Other improvements obtained at different number of OFDMA

Fig. 24. BER Comparison between Turbo code, CPC at 16QAM , N=128.

Fig. 25. BER Comparison between Turbo code, CPC at 16 QAM, N=512.

The comparison for modulation type 64QAM between modified CPC and convolutional code, LDPC code and turbo code are shown through Figure 26 to Figure 31 through the

subcarriers (512) as shown from Figure 25.

The results of comparisons between CPC and LDPC code through the fading channel are obtained from Figure 22 to Figure 23. These comparisons are obtained for modulation 16QAQM at different SNR values. From Figure 22, it is shown that SNR is improved by about 2.5 dB at BER equals 10-2 for the number of subcarriers equals 128. Another improvement is also obtained at different number of OFDM subcarriers (512) as shown in Figure 23.

Fig. 22. BER Comparison between LDPC code, CPC at 16 QAM, N=128.

Fig. 23. BER Comparison between LDPC code, CPC at 16 QAM, N=512.

The results of comparisons between modified CPC and turbo code through the fading channel are shown from Figure 24 to Figure 25. These comparisons are done at different

The results of comparisons between CPC and LDPC code through the fading channel are obtained from Figure 22 to Figure 23. These comparisons are obtained for modulation 16QAQM at different SNR values. From Figure 22, it is shown that SNR is improved by about 2.5 dB at BER equals 10-2 for the number of subcarriers equals 128. Another improvement is also obtained at different number of OFDM subcarriers (512) as shown in

Fig. 22. BER Comparison between LDPC code, CPC at 16 QAM, N=128.

Fig. 23. BER Comparison between LDPC code, CPC at 16 QAM, N=512.

The results of comparisons between modified CPC and turbo code through the fading channel are shown from Figure 24 to Figure 25. These comparisons are done at different

Figure 23.

SNR values for modulation type 16QAM. There is an improvement in SNR by more than 8 dB at BER equals 10-2 for 16QAM and number of OFDMA subcarriers equals 128, this is shown from Figure 24. Other improvements obtained at different number of OFDMA subcarriers (512) as shown from Figure 25.

Fig. 24. BER Comparison between Turbo code, CPC at 16QAM , N=128.

Fig. 25. BER Comparison between Turbo code, CPC at 16 QAM, N=512.

The comparison for modulation type 64QAM between modified CPC and convolutional code, LDPC code and turbo code are shown through Figure 26 to Figure 31 through the

Performance Evaluation of WiMAX System Using Different Coding Techniques 283

can be obtained when the number of subcarriers is increased to 512 as shown in Figure 27. Figure 28 and Figure 29 show the results of comparisons between modified CPC and LDPC code through the fading channel are shown at different received SNR values. We conclude that modified CPC gives good results at different SNR. Figure 28 shows that modified CPC coding technique gives better results than LDPC coding technique at 64QAM and OFDM subcarriers equal 128. The improvement is more than 1.5 dB for 10 -2. Also, there will be an improvement, when the number of subcarriers is increased to 512 as shown in Figure 29.

Fig. 28. BER Comparison between LDPC code, CPC at 64 QAM, N=128.

Fig. 29. BER Comparison between LDPC code, CPC at 64 QAM, N=512.

fading channel.. BER versus different received SNR values are shown in these figures. These comparisons are obtained and number of subcarriers equals 128 and 512 respectively. The comparisons between modified CPC and convolutional code through the fading channel are shown in both Figure 26 and Figure 27.

Fig. 26. BER Comparison between Conv code, CPC at 64QAM, N=128.

Fig. 27. BER Comparison between Conv. code, CPC at 64 QAM, N=512.

From Figure 26, it is shown that SNR will be improved by more than 2 dB at BER equals 10 -2 for modulation type 64QAM and number of subcarriers equals 128. Also, an improvement

fading channel.. BER versus different received SNR values are shown in these figures. These comparisons are obtained and number of subcarriers equals 128 and 512 respectively. The comparisons between modified CPC and convolutional code through the fading channel are

Fig. 26. BER Comparison between Conv code, CPC at 64QAM, N=128.

Fig. 27. BER Comparison between Conv. code, CPC at 64 QAM, N=512.

From Figure 26, it is shown that SNR will be improved by more than 2 dB at BER equals 10 -2 for modulation type 64QAM and number of subcarriers equals 128. Also, an improvement

shown in both Figure 26 and Figure 27.

can be obtained when the number of subcarriers is increased to 512 as shown in Figure 27. Figure 28 and Figure 29 show the results of comparisons between modified CPC and LDPC code through the fading channel are shown at different received SNR values. We conclude that modified CPC gives good results at different SNR. Figure 28 shows that modified CPC coding technique gives better results than LDPC coding technique at 64QAM and OFDM subcarriers equal 128. The improvement is more than 1.5 dB for 10 -2. Also, there will be an improvement, when the number of subcarriers is increased to 512 as shown in Figure 29.

Fig. 28. BER Comparison between LDPC code, CPC at 64 QAM, N=128.

Fig. 29. BER Comparison between LDPC code, CPC at 64 QAM, N=512.

Performance Evaluation of WiMAX System Using Different Coding Techniques 285

Due to the lower code rate of CPC (1/4), better results should be obtained comparing to the other coding techniques (Convolution – Turbo – LDPC) which has higher coding rate (1/2 or 1/3). But from the view of complexity, CPC is less complex than turbo or LDPC, so the type of coding can be used as an optional code instead of turbo or LDPC types. Another technique can be used with modified CPC to modify or increase its code rate, this technique

To modify the rate of the coding process, a puncture technique is used. This technique enables to have a code rate equals 1/3. The modification is done by applying the puncture to the columns only, resulting in code rate of 2/3. So the overall code rate will be (2/3) x (1/2) = (1/3). Puncture enables to reduce the redundancy bits but on other hand it leads to increase the BER. From Figure 32 to Figure 42 the result of using CPC with puncture is shown, through AWGN plus fading channel, comparing with convolutional code, turbo code and LDPC code. From Figure 32 to Figure 34 BER versus different received SNR values are shown for the comparison between modified CPC, puncture CPC and convolutional code, LDPC code and turbo code respectively through the fading channel. These comparisons are obtained for modulation type 16QAM and number of subcarriers equals

Fig. 32. BER Comparison between Conv code, CPC, punctured CPC at 16 QAM, N=128.

called *puncture technique*.

**9. Punctured CPC** 

128.

From Figure 30 to Figure 31, the results produced from the comparisons between modified CPC and turbo code are shown at different received SNR values for modulation 64 QAM. As shown from these figures, modified CPC method gives good results compared to turbo coding. Figure 30 shows that using modified CPC method can give better results than turbo coding technique at 64QAM and OFDM subcarriers equals 128. This improvement is about than 2 dB for 10 -2. Also other improvements can be obtained at different number of OFDM subcarriers (512) as shown in Figure 31.

Fig. 30. BER Comparison between Turbo code, CPC at 64 QAM, N=128.

Fig. 31. BER Comparison between Turbo code, CPC at 64 QAM, N=512.

Due to the lower code rate of CPC (1/4), better results should be obtained comparing to the other coding techniques (Convolution – Turbo – LDPC) which has higher coding rate (1/2 or 1/3). But from the view of complexity, CPC is less complex than turbo or LDPC, so the type of coding can be used as an optional code instead of turbo or LDPC types. Another technique can be used with modified CPC to modify or increase its code rate, this technique called *puncture technique*.

#### **9. Punctured CPC**

284 Advanced Transmission Techniques in WiMAX

From Figure 30 to Figure 31, the results produced from the comparisons between modified CPC and turbo code are shown at different received SNR values for modulation 64 QAM. As shown from these figures, modified CPC method gives good results compared to turbo coding. Figure 30 shows that using modified CPC method can give better results than turbo coding technique at 64QAM and OFDM subcarriers equals 128. This improvement is about than 2 dB for 10 -2. Also other improvements can be obtained at different number of OFDM

Fig. 30. BER Comparison between Turbo code, CPC at 64 QAM, N=128.

Fig. 31. BER Comparison between Turbo code, CPC at 64 QAM, N=512.

subcarriers (512) as shown in Figure 31.

To modify the rate of the coding process, a puncture technique is used. This technique enables to have a code rate equals 1/3. The modification is done by applying the puncture to the columns only, resulting in code rate of 2/3. So the overall code rate will be (2/3) x (1/2) = (1/3). Puncture enables to reduce the redundancy bits but on other hand it leads to increase the BER. From Figure 32 to Figure 42 the result of using CPC with puncture is shown, through AWGN plus fading channel, comparing with convolutional code, turbo code and LDPC code. From Figure 32 to Figure 34 BER versus different received SNR values are shown for the comparison between modified CPC, puncture CPC and convolutional code, LDPC code and turbo code respectively through the fading channel. These comparisons are obtained for modulation type 16QAM and number of subcarriers equals 128.

Fig. 32. BER Comparison between Conv code, CPC, punctured CPC at 16 QAM, N=128.

Performance Evaluation of WiMAX System Using Different Coding Techniques 287

Fig. 35. BER Comparison between Conv code, CPC, punctured CPC at 16 QAM, N=512.

Fig. 36. BER Comparison between LDPC code, CPC, punctured CPC at 16 QAM, N=512.

Fig. 33. BER Comparison between LDPC code, CPC, punctured CPC at 16 QAM, N=128.

Fig. 34. BER Comparison between Turbo code, CPC, punctured CPC at 16 QAM, N=128.

From Figure 32 it is shown that the results obtained from LDPC code is approximately the same as puncture modified CPC, but LDPC is more complicated than the proposed method. From Figure 35 to Figure 37 BER versus different received SNR values are shown for the comparison between modified CPC, puncture CPC and convolutional code, LDPC code and turbo code respectively through the fading channel. These comparisons are obtained for modulation type 16QAM and number of subcarriers equals 512.

Fig. 33. BER Comparison between LDPC code, CPC, punctured CPC at 16 QAM, N=128.

Fig. 34. BER Comparison between Turbo code, CPC, punctured CPC at 16 QAM, N=128.

modulation type 16QAM and number of subcarriers equals 512.

From Figure 32 it is shown that the results obtained from LDPC code is approximately the same as puncture modified CPC, but LDPC is more complicated than the proposed method. From Figure 35 to Figure 37 BER versus different received SNR values are shown for the comparison between modified CPC, puncture CPC and convolutional code, LDPC code and turbo code respectively through the fading channel. These comparisons are obtained for

Fig. 35. BER Comparison between Conv code, CPC, punctured CPC at 16 QAM, N=512.

Fig. 36. BER Comparison between LDPC code, CPC, punctured CPC at 16 QAM, N=512.

Performance Evaluation of WiMAX System Using Different Coding Techniques 289

Fig. 39. BER Comparison between LDPC code, CPC, punctured CPC at 64 QAM, N=128.

Fig. 40. BER Comparison between Turbo code, CPC, punctured CPC at 64 QAM, N=128.

obtained for modulation type 64QAM and number of subcarriers equals 512.

From Figure 39 it is shown that the results obtained from LDPC code is better than t puncture modified CPC until SNR = 16db, but LDPC is more complicated than the proposed method. From Figure 35 to Figure 37 BER versus different received SNR values are shown for the comparison between modified CPC, puncture CPC and convolutional code, LDPC code and turbo code respectively through the fading channel. These comparisons are

Fig. 37. BER Comparison between Turbo code, CPC, punctured CPC at 16 QAM, N=512

From Figure 36 it is shown that the results obtained from LDPC code is better than puncture modified CPC until SNR = 12 db, but LDPC is more complicated than the proposed method. From Figure 35 to Figure 37 BER versus different received SNR values are shown for the comparison between modified CPC, puncture CPC and convolutional code, LDPC code and turbo code respectively through the fading channel. These comparisons are obtained for modulation type 64QAM and number of subcarriers equals 128.

Fig. 38. BER Comparison between Conv code, CPC, punctured CPC at 64 QAM, N=128.

Fig. 37. BER Comparison between Turbo code, CPC, punctured CPC at 16 QAM, N=512

Fig. 38. BER Comparison between Conv code, CPC, punctured CPC at 64 QAM, N=128.

modulation type 64QAM and number of subcarriers equals 128.

From Figure 36 it is shown that the results obtained from LDPC code is better than puncture modified CPC until SNR = 12 db, but LDPC is more complicated than the proposed method. From Figure 35 to Figure 37 BER versus different received SNR values are shown for the comparison between modified CPC, puncture CPC and convolutional code, LDPC code and turbo code respectively through the fading channel. These comparisons are obtained for

Fig. 39. BER Comparison between LDPC code, CPC, punctured CPC at 64 QAM, N=128.

From Figure 39 it is shown that the results obtained from LDPC code is better than t puncture modified CPC until SNR = 16db, but LDPC is more complicated than the proposed method. From Figure 35 to Figure 37 BER versus different received SNR values are shown for the comparison between modified CPC, puncture CPC and convolutional code, LDPC code and turbo code respectively through the fading channel. These comparisons are obtained for modulation type 64QAM and number of subcarriers equals 512.

Fig. 40. BER Comparison between Turbo code, CPC, punctured CPC at 64 QAM, N=128.

Performance Evaluation of WiMAX System Using Different Coding Techniques 291

From Figure 42 and Figure 43 it is shown that the results obtained from LDPC code and turbo code is better than puncture modified CPC, but LDPC and turbo code is more

Fig. 43. BER Comparison between Turbo code, CPC, punctured CPC at 64 QAM, N=512.

complicated than the proposed method.

Fig. 41. BER Comparison between Conv code, CPC, punctured CPC at 64 QAM, N=512.

Fig. 42. BER Comparison between LDPC code, CPC, punctured CPC at 64 QAM, N=512.

Fig. 41. BER Comparison between Conv code, CPC, punctured CPC at 64 QAM, N=512.

Fig. 42. BER Comparison between LDPC code, CPC, punctured CPC at 64 QAM, N=512.

From Figure 42 and Figure 43 it is shown that the results obtained from LDPC code and turbo code is better than puncture modified CPC, but LDPC and turbo code is more complicated than the proposed method.

Fig. 43. BER Comparison between Turbo code, CPC, punctured CPC at 64 QAM, N=512.

Performance Evaluation of WiMAX System Using Different Coding Techniques 293

The receiver, by using one received antenna, chooses the best received signal according to its

In this chapter, we explained CPC method as a coding technique and our modification for it. Also the implementation of CPC in WiMAX system and the comparisons between its results and the results of other coding techniques such as convolutional, turbo and LDPC are investigated at different SNR for different number of subcarriers and at different types of

[1] Nam Yul Yu, Young Kim, and Pil Joong Lee (2000). Iterative Decoding of Product Codes

[2] R. M. Pyndiah, (1998). Near-Optimum Decoding of Product Codes: Block Turbo Codes,"

[3] T. Shohon, Y. Soutome and H. Ogiwara. (1999). Simple Computation Method of Soft

[4] Omar Aitsab and Ramesh Pyndiah. (1996). Performance of Reed Solomon Block

[5] David Rankin and T. Aaron Gulliver. (2001). Single Parity Check Product Codes," *IEEE* 

[6] *Lin, S. and Costello. (1983*). Error Control Coding: Fundamentals and Applications*,* 

[7] V. Tarokh V, N. Seshadri, and A. R. Calderbank. (1998). Space-Time Codes for High Data

[8] Orhan Gazi and Ali Özgür Ylmaz (2006). Turbo Product Codes Based on Convolutional

*Communications (ISCC 2000), France, pp. 732 -737, July.* 

*Trans. Commun., vol. 49, no. 8, pp. 1354-1362, Aug.* 

Codes," *ETRI Journal, Volume 28, Number 4, Aug.* 

*IEEE Trans. Inform. Theory,* vol. 44, no. 2, pp. 744-765, *Mar.*

*Prentice Hall, Englewood Cliffs, New Jersey.* 

*IEEE Trans. Communication., vol. 46, no. 8, pp. 1003-1010, Aug.* 

*EIC Trans. Fundamentals, vol. E82-A, no. 10, pp. 2199–2203, Oct.* 

Composed of Extended Hamming Codes," *5th IEEE Symposium on Computers and* 

Value for Iterative Decoding of Product Code Composed of Linear Block Code,"

Turbo Codes," *Proc. IEEE LOBECOM' 96 Conf., London, U.K., vol. 1/3, pp. 121-*

Rate Wireless Communication: Performance Criterion and Code Construction,"

Fig. 45. Modified Transmit Delay Diversity Scheme

highest power.

**11. Conclusions** 

**12. References** 

modulation (16QAM – 64QAM).

*125, Nov.* 

#### **10. Delay diversity scheme**

In this section, both puncture and Delay Diversity Scheme (DDS) are together used to increase the efficiency of CPC system. To further improve the diversity of the channels a transmit diversity technique may be utilized.

Many transmit diversity techniques have been explored. One such technique is the transmit delay diversity Figure 6.44. In transmit delay diversity a transmitter utilizes two antennas that transmit the same signal, with the second antenna transmitting a delayed replica of that transmitted by the first antenna. By so doing, the second antenna creates diversity by establishing a second set of independent multipath elements that may be collected at the receiver.

Fig. 44. Transmit Delay Diversity Scheme

If the multipath generated by the first transmitter fades, the multipath generated by the second transmitter may not, in which case an acceptable SNR will be maintained at the receiver. This technique is easy to implement, because only the composite TX0+TX1 channel is estimated at the receiver. The biggest drawback to transmit delay diversity is that it increases the effective delay spread of the channel, and can perform poorly when the multipath introduced by the second antenna falls upon, and interacts destructively with, the multipath of the first antenna, thereby reducing the overall level of diversity.

Our simulator for delay diversity technique is based on passing the same signal through the same path during two time intervals by using only one transmitted antenna Figure 6.45, not two transmitted antennas as in delay diversity technique. During these intervals the channel will have different fading and AWGN characteristics over the time.

Fig. 45. Modified Transmit Delay Diversity Scheme

The receiver, by using one received antenna, chooses the best received signal according to its highest power.

#### **11. Conclusions**

292 Advanced Transmission Techniques in WiMAX

In this section, both puncture and Delay Diversity Scheme (DDS) are together used to increase the efficiency of CPC system. To further improve the diversity of the channels a

Many transmit diversity techniques have been explored. One such technique is the transmit delay diversity Figure 6.44. In transmit delay diversity a transmitter utilizes two antennas that transmit the same signal, with the second antenna transmitting a delayed replica of that transmitted by the first antenna. By so doing, the second antenna creates diversity by establishing

If the multipath generated by the first transmitter fades, the multipath generated by the second transmitter may not, in which case an acceptable SNR will be maintained at the receiver. This technique is easy to implement, because only the composite TX0+TX1 channel is estimated at the receiver. The biggest drawback to transmit delay diversity is that it increases the effective delay spread of the channel, and can perform poorly when the multipath introduced by the second antenna falls upon, and interacts destructively with, the

Our simulator for delay diversity technique is based on passing the same signal through the same path during two time intervals by using only one transmitted antenna Figure 6.45, not two transmitted antennas as in delay diversity technique. During these intervals the channel

multipath of the first antenna, thereby reducing the overall level of diversity.

will have different fading and AWGN characteristics over the time.

a second set of independent multipath elements that may be collected at the receiver.

**10. Delay diversity scheme** 

transmit diversity technique may be utilized.

Fig. 44. Transmit Delay Diversity Scheme

In this chapter, we explained CPC method as a coding technique and our modification for it. Also the implementation of CPC in WiMAX system and the comparisons between its results and the results of other coding techniques such as convolutional, turbo and LDPC are investigated at different SNR for different number of subcarriers and at different types of modulation (16QAM – 64QAM).

#### **12. References**


**Part 3** 

**Mobile WiMAX Techniques and** 

**Interconnection with Other Technologies** 


### **Part 3**

### **Mobile WiMAX Techniques and Interconnection with Other Technologies**

294 Advanced Transmission Techniques in WiMAX

[9] Shannon, C. E. (1948). A mathematical theory of Communication," *Bell Syst. Tech. J., vol.* 

[10] Proakis, J. G. and Salehi, M. (1993). Communication Systems Engineering, *Prentice Hall,* 

[11] K.Miyauchi, S.Seki, and H. Ishio (1976). New Techniques for Generating and Detecting

Multilevel Signal Formats*," IEEE Trans Communication, vol. COM-24, pp. 263-267,* 

*27, pp. 379–423, 623–656, July and Oct.* 

*New Jersey.* 

*Feb.* 

**14** 

*France* 

**802.16e & 802.11s** 

*ARTIMIA, Malakoff* 

Tarek Bchini and Mina Ouabiba

**Interaction and Interconnection Between** 

With the rapid evolution of wireless and mobile networks, and the emergence of several standards that use different technologies, the problem of compatibility between these technologies, or the transition from one technology to the other by the mobile station without interruption of services, becomes a real challenge to face, to ensure a good Quality

In this context, we will analyze Mobile Stream Control Transport Protocol (MSCTP) and IEEE 802.21 technology as two vertical handover mechanisms between two of mobile networks: IEEE 802.11s, and mobile WIMAX. The simulations will be run under Network

The mobile WiMAX (IEEE 802.16e) [2] is a mobile extension of the IEEE 802.16 standard [3]. IEEE 802.16 defines the specifications for radio metropolitan networks or WMAN (Wireless Metropolitan Area Network), offering broadband to achieve a high flow rate and using

The IEEE 802.16e is suitable for any kind of traffic thanks to its flexibility justified by its

There are two kinds of Handover in the mobile WiMAX: Intra-ASN Handover (layer 2: no change of IP address) and Inter-ASN Handover (layer 3: IP address change) [4] [5]. For Intra-ASN, two mechanisms have been specified: Hard handover for the low speed and Soft handover for the high speed; and for Inter-ASN Handover it defines: Mobile IPv4 or Client-

The architecture of 802.16e is composed of mobile stations (MS), that communicate freely (radio link) with base stations (BS), which act as an intermediates gateways with the terrestrial infrastructure of IP network. The base stations themselves are connected to the network elements called ASN-GW (gateways) which manages their connection with the IP

**1. Introduction 1.1 Problematic** 

of Service (QoS) for the client.

**1.2 Mobile Wimax (IEEE 802.16e)** 

techniques to cover large areas [3].

MIPv4, and Proxy-MIPv4 [4].

network [4] [5].

three MAC layers [2] and its use of IP protocol.

Simulator 2 (NS2) [1].

### **Interaction and Interconnection Between 802.16e & 802.11s**

Tarek Bchini and Mina Ouabiba *ARTIMIA, Malakoff France* 

#### **1. Introduction**

#### **1.1 Problematic**

With the rapid evolution of wireless and mobile networks, and the emergence of several standards that use different technologies, the problem of compatibility between these technologies, or the transition from one technology to the other by the mobile station without interruption of services, becomes a real challenge to face, to ensure a good Quality of Service (QoS) for the client.

In this context, we will analyze Mobile Stream Control Transport Protocol (MSCTP) and IEEE 802.21 technology as two vertical handover mechanisms between two of mobile networks: IEEE 802.11s, and mobile WIMAX. The simulations will be run under Network Simulator 2 (NS2) [1].

#### **1.2 Mobile Wimax (IEEE 802.16e)**

The mobile WiMAX (IEEE 802.16e) [2] is a mobile extension of the IEEE 802.16 standard [3].

IEEE 802.16 defines the specifications for radio metropolitan networks or WMAN (Wireless Metropolitan Area Network), offering broadband to achieve a high flow rate and using techniques to cover large areas [3].

The IEEE 802.16e is suitable for any kind of traffic thanks to its flexibility justified by its three MAC layers [2] and its use of IP protocol.

There are two kinds of Handover in the mobile WiMAX: Intra-ASN Handover (layer 2: no change of IP address) and Inter-ASN Handover (layer 3: IP address change) [4] [5]. For Intra-ASN, two mechanisms have been specified: Hard handover for the low speed and Soft handover for the high speed; and for Inter-ASN Handover it defines: Mobile IPv4 or Client-MIPv4, and Proxy-MIPv4 [4].

The architecture of 802.16e is composed of mobile stations (MS), that communicate freely (radio link) with base stations (BS), which act as an intermediates gateways with the terrestrial infrastructure of IP network. The base stations themselves are connected to the network elements called ASN-GW (gateways) which manages their connection with the IP network [4] [5].

Interaction and Interconnection Between 802.16e & 802.11s 299

 Mesh Point (MP) which supports (fully or partially) mesh relay functions, and implement operations such as channel selection, neighbor discovery, and forming and association with neighbors. Additionally, MPs communicate with their neighbors and

 Mesh Access Point (MAP= MP+AP) which is a MP but acts as an AP as well. Therefore, MAPs can operate in a WLAM Mesh or as part of legacy IEEE 802.11 modes. Mesh Portal (MPP=MP+Bridge) is another kind of MP that allows the interconnection of multiple WLAN meshes to form a network of mesh networks. Moreover, MPP can function as bridges or gateways to connect to other wired or wireless networks in the DS.

In our work, we analyze two vertical handover mechanisms that will be used between the mobile WiMAX and the IEEE 802.11s, and in this section we will present the two proposed

The transport layer mobility is proposed as an alternative to the network layer mobility to support integrated mobility. The management of mobility in the transport layer is made exclusively by Stream Control Transmission Protocol (SCTP) [8] and its extension: Dynamic

Simple Station (STA): outside of the WLAN Mesh, connected via Mesh AP.

The architecture of IEEE 802.11s is presented in the figure 2 below:

forward traffic on behalf of other MPs.

Fig. 2. IEEE 802.11s Architecture

Address Reconfiguration (DAR) [9].

mechanisms.

**2.1 MSCTP protocol** 

**2. Vertical handover mechanisms proposed** 

The NAP (Network Access Provider) is an entity that provides the infrastructure for radio access to one or more providers of network services. It can control one or more ASN (Access Service Network) which is composed of one or more BS and one or more gateways.

The NSP (Network Service provider) is an entity that provides IP connectivity and network services to subscribers compatible with the level of service it establishes with subscribers. An NSP may also establish roaming agreements with other providers of network services and contractual agreements with third-party providers of application (for example, ASP: Application Service Provider) to provide IP services to subscribers.

A NSP control one or more CSN (Connectivity Service Network) which is the core of the WiMAX network.

The architecture of mobile WiMAX is presented in the figure 1:

Fig. 1. IEEE 802.16e Architecture [4] [5]

#### **1.3 IEEE 802.11s**

IEEE 802.11s [6] [7] is an amendment being developed to the IEEE 802.11 WLAN standard, and aims to implement mobility on Ad-Hoc networks with acceptable debit.

In September 2003, IEEE formed the 802.11s SG which, in July 2004, became the "Extended Service Set (ESS) Mesh Networking" or 802.11s Task Group (TGs), and it is the most advanced group of the 802.11 WG.

The current objective of this TG is to apply mesh technology to WLANs by defining a Wireless Distribution System (WDS) used to build a wireless infrastructure with MAC-layer broadcast/multicast support in addition to the unicast transmissions. The TG should produce a protocol that specifies the installation, configuration, and operation of WLAN mesh. Moreover, the specification should include the extensions in topology formation to make the WLAN mesh self-configure and self-organized, and support for multi-channel, and multi-radio devices. At the MAC layer, a selection path protocol should be incorporated, instead of assigning the routing task to the network layer.

The WLAN Mesh architecture comprises the following IEEE 802.11 based elements:

The NAP (Network Access Provider) is an entity that provides the infrastructure for radio access to one or more providers of network services. It can control one or more ASN (Access

The NSP (Network Service provider) is an entity that provides IP connectivity and network services to subscribers compatible with the level of service it establishes with subscribers. An NSP may also establish roaming agreements with other providers of network services and contractual agreements with third-party providers of application (for example, ASP:

A NSP control one or more CSN (Connectivity Service Network) which is the core of the

IEEE 802.11s [6] [7] is an amendment being developed to the IEEE 802.11 WLAN standard,

In September 2003, IEEE formed the 802.11s SG which, in July 2004, became the "Extended Service Set (ESS) Mesh Networking" or 802.11s Task Group (TGs), and it is the most

The current objective of this TG is to apply mesh technology to WLANs by defining a Wireless Distribution System (WDS) used to build a wireless infrastructure with MAC-layer broadcast/multicast support in addition to the unicast transmissions. The TG should produce a protocol that specifies the installation, configuration, and operation of WLAN mesh. Moreover, the specification should include the extensions in topology formation to make the WLAN mesh self-configure and self-organized, and support for multi-channel, and multi-radio devices. At the MAC layer, a selection path protocol should be

and aims to implement mobility on Ad-Hoc networks with acceptable debit.

incorporated, instead of assigning the routing task to the network layer.

The WLAN Mesh architecture comprises the following IEEE 802.11 based elements:

Service Network) which is composed of one or more BS and one or more gateways.

Application Service Provider) to provide IP services to subscribers.

The architecture of mobile WiMAX is presented in the figure 1:

WiMAX network.

**1.3 IEEE 802.11s** 

Fig. 1. IEEE 802.16e Architecture [4] [5]

advanced group of the 802.11 WG.


The architecture of IEEE 802.11s is presented in the figure 2 below:

Fig. 2. IEEE 802.11s Architecture

#### **2. Vertical handover mechanisms proposed**

In our work, we analyze two vertical handover mechanisms that will be used between the mobile WiMAX and the IEEE 802.11s, and in this section we will present the two proposed mechanisms.

#### **2.1 MSCTP protocol**

The transport layer mobility is proposed as an alternative to the network layer mobility to support integrated mobility. The management of mobility in the transport layer is made exclusively by Stream Control Transmission Protocol (SCTP) [8] and its extension: Dynamic Address Reconfiguration (DAR) [9].

Interaction and Interconnection Between 802.16e & 802.11s 301

The new primary IP address is now the second IP address obtained. The CN sends at this moment all the messages to the new IP address of MS via the foreign network. And finally, when the MS leave definitively the coverage area of the home network, it informs the CN to delete the first IP address of the association, which the CN confirms with an ACK [10] [11] [12]. To use MSCTP, the only requirement is that the both endpoints should implement

By applying the MSCTP in the case of vertical handover between 802.16e and 802.11s, the protocol will use the multi-homing technique to open two IP sessions with the BS of the mobile WiMAX network and the AP of the Wireless mesh network to avoid the service

IEEE 802.21 or Media Independent Handover (MIH) [13] [14] [15] is a recent evolution for all networks, that providing capabilities to detect and initiate handover from one network to another. It designed a new function to control access to the lower layers (Layers 1 and 2). This new function provides new service access points (SAPs) and allows the information to be queried by the upper layers (Layer 3 and higher). Both mobile device and network hardware must implement the standard to work, but everything should remain backward

The standard allows simply to provide information that help to the initiation of handover, the selection of the network and the activation of the interface. The execution and the

In MIH Function (MIHF), there are three services that allow the passage of messages along the stack. The table 1 below compiled from IEEE 802.21, outlines the basic functions of these

Origin Destination Use cases

MIHF or upper layer Remote or local stack

MIHF or lower layer Remote or local stack

Upper or lower layer Remote or local stack Link up/down/going down, transmissions

switch links, get status

information elements (IEs), neighbor reports

status

MSCTP protocol.

interruption during handover.

compatible for non-MIH aware devices.

Event MIHF or lower layer

Command MIHF or upper layer

Information Upper or lower layer

Table 1. MIH services

port

Secure or insecure

The MIH architecture is illustrated in the figure 4 below:

decision of handover is not part of the standard.

**2.2 IEEE 802.21 or MIH** 

services [14].

MIH services

SCTP extended with DAR constitute Mobile SCTP (MSCTP) [10] [11] [12].

MSCTP was designed in order to avoid the connection disruptions observed with TCP or UDP during a change of IP address. It is a transport layer protocol similar to Transmission Control Protocol (TCP). It provides point-to-point communication oriented connection between applications running on different hosts. The major difference with TCP is the multihoming; it allows by multi-homing to manage multiple IP addresses in terminal nodes by conserving the point-to-point connection intact (see figure 3).

Fig. 3. MSCTP vs TCP and protocol stack

In the beginning of the communication between a mobile station (MS) and its correspondent (CN) implementing both MSCTP protocol; in the MS, there is only one IP address chosen as primary address, and used as destination address for the current transmission. The other IP addresses are used only for retransmissions. The DAR extension allows to MS to add, delete and change IP addresses during a SCTP session, without affecting the connection established, by using address configuration messages.

During the communication with the CN; when the MS changes from its home network to a foreign network passing by handover Area (at the beginning of the coverage area of foreign network), it receives an IP address from the foreign network either by contacting a DHCP, or by automatic configuration of IPv4 address. The MS is now able to establish other link with its CN through this second IP address obtained, and may become accessible via the foreign network. Then it sends its second IP address via the home network to its CN, and the CN will add the new IP address to the association identifying the connection with the MS and sends an ACK to the MS to confirm. After, when the MS begin to leave the coverage area of its home network to the coverage area of the foreign network, it notifies the CN to assign the new IP address as primary IP address, which the CN approves with an ACK.

The new primary IP address is now the second IP address obtained. The CN sends at this moment all the messages to the new IP address of MS via the foreign network. And finally, when the MS leave definitively the coverage area of the home network, it informs the CN to delete the first IP address of the association, which the CN confirms with an ACK [10] [11] [12]. To use MSCTP, the only requirement is that the both endpoints should implement MSCTP protocol.

By applying the MSCTP in the case of vertical handover between 802.16e and 802.11s, the protocol will use the multi-homing technique to open two IP sessions with the BS of the mobile WiMAX network and the AP of the Wireless mesh network to avoid the service interruption during handover.

#### **2.2 IEEE 802.21 or MIH**

300 Advanced Transmission Techniques in WiMAX

MSCTP was designed in order to avoid the connection disruptions observed with TCP or UDP during a change of IP address. It is a transport layer protocol similar to Transmission Control Protocol (TCP). It provides point-to-point communication oriented connection between applications running on different hosts. The major difference with TCP is the multihoming; it allows by multi-homing to manage multiple IP addresses in terminal nodes by

In the beginning of the communication between a mobile station (MS) and its correspondent (CN) implementing both MSCTP protocol; in the MS, there is only one IP address chosen as primary address, and used as destination address for the current transmission. The other IP addresses are used only for retransmissions. The DAR extension allows to MS to add, delete and change IP addresses during a SCTP session, without affecting the connection

During the communication with the CN; when the MS changes from its home network to a foreign network passing by handover Area (at the beginning of the coverage area of foreign network), it receives an IP address from the foreign network either by contacting a DHCP, or by automatic configuration of IPv4 address. The MS is now able to establish other link with its CN through this second IP address obtained, and may become accessible via the foreign network. Then it sends its second IP address via the home network to its CN, and the CN will add the new IP address to the association identifying the connection with the MS and sends an ACK to the MS to confirm. After, when the MS begin to leave the coverage area of its home network to the coverage area of the foreign network, it notifies the CN to assign the new IP address as primary IP address, which the CN approves with an ACK.

SCTP extended with DAR constitute Mobile SCTP (MSCTP) [10] [11] [12].

conserving the point-to-point connection intact (see figure 3).

Fig. 3. MSCTP vs TCP and protocol stack

established, by using address configuration messages.

IEEE 802.21 or Media Independent Handover (MIH) [13] [14] [15] is a recent evolution for all networks, that providing capabilities to detect and initiate handover from one network to another. It designed a new function to control access to the lower layers (Layers 1 and 2). This new function provides new service access points (SAPs) and allows the information to be queried by the upper layers (Layer 3 and higher). Both mobile device and network hardware must implement the standard to work, but everything should remain backward compatible for non-MIH aware devices.

The standard allows simply to provide information that help to the initiation of handover, the selection of the network and the activation of the interface. The execution and the decision of handover is not part of the standard.

In MIH Function (MIHF), there are three services that allow the passage of messages along the stack. The table 1 below compiled from IEEE 802.21, outlines the basic functions of these services [14].


Table 1. MIH services

The MIH architecture is illustrated in the figure 4 below:

Interaction and Interconnection Between 802.16e & 802.11s 303

Based on common model proposed in the section before, with MSCTP protocol, the end users (the MS moving between the two cells and its correspondent node or server) must

The architecture of simulated model proposed for the interconnection between IEEE 802.16e

The exchange of information between the MS and its correspondent node (CN) during the MS mobility scenario between IEEE 802.16e and IEEE 802.11s using MSCTP is illustrated in

Fig. 6. Messages exchanged during simulation of the protocol MSCTP

**3.2 Simulation model based on MSCTP protocol** 

and IEEE 802.11s using MSCTP is illustrated in the figure 5 below:

implement the MSCTP protocol.

Fig. 5. Interconnection model using MSCTP

the figure 6 below:

Fig. 4. MIH architecture

#### **3. Interconnection models proposed**

#### **3.1 Common simulation model and common scenario**

We describe in this section the common interconnection model proposed between IEEE 802.16e and IEEE 802.11s using MSCTP or IEEE 802.21 during the vertical handover; and the MS mobility scenarios between the two networks. We will assume an 802.16e cell with coverage of 1 km radius and an 802.11s cell with coverage of 300 m radius (these choices of radius values are based on the nature of the test environment that is an urban area which is not very dense). And, the two cells have a common area (handover area) with a variable and maximum distance of 180 m between both limits of cells (this choice of the surface of common area between cells is based on the time needed for the handover simulation).

The Base station (BS) of 802.16e network is linked to an ASN-GW that is linked too via IP network to a CSN (WiMAX ISP); the Access Point (AP) of 802.11s network is linked to a router that is linked too via IP network to a WIFI CSN; and the two CSNs are connected together and with the distant servers via Internet network.

We will evaluate the mechanisms through two mobility scenarios for the simulations: the case where the mobile move from 802.16e cell to 802.11s cell, and the opposite case. In the two scenarios, the mobile station (MS) traverses 200 m in 802.16e or 802.11s cell, and traverses in handover area (common area between 802.16e and 802.11s cells) 100 m.

We will propose three MS speeds to see the impact of speed increasing on the handover. So, we will propose: 5 m/s = 18 km/h; 10 m/s = 36 km/h; and 20 m/s = 72 km/h as mobile speeds for the simulations.

#### **3.2 Simulation model based on MSCTP protocol**

302 Advanced Transmission Techniques in WiMAX

We describe in this section the common interconnection model proposed between IEEE 802.16e and IEEE 802.11s using MSCTP or IEEE 802.21 during the vertical handover; and the MS mobility scenarios between the two networks. We will assume an 802.16e cell with coverage of 1 km radius and an 802.11s cell with coverage of 300 m radius (these choices of radius values are based on the nature of the test environment that is an urban area which is not very dense). And, the two cells have a common area (handover area) with a variable and maximum distance of 180 m between both limits of cells (this choice of the surface of common area between cells is based on the time needed for the handover simulation).

The Base station (BS) of 802.16e network is linked to an ASN-GW that is linked too via IP network to a CSN (WiMAX ISP); the Access Point (AP) of 802.11s network is linked to a router that is linked too via IP network to a WIFI CSN; and the two CSNs are connected

We will evaluate the mechanisms through two mobility scenarios for the simulations: the case where the mobile move from 802.16e cell to 802.11s cell, and the opposite case. In the two scenarios, the mobile station (MS) traverses 200 m in 802.16e or 802.11s cell, and

We will propose three MS speeds to see the impact of speed increasing on the handover. So, we will propose: 5 m/s = 18 km/h; 10 m/s = 36 km/h; and 20 m/s = 72 km/h as mobile

traverses in handover area (common area between 802.16e and 802.11s cells) 100 m.

Fig. 4. MIH architecture

speeds for the simulations.

**3. Interconnection models proposed** 

**3.1 Common simulation model and common scenario** 

together and with the distant servers via Internet network.

Based on common model proposed in the section before, with MSCTP protocol, the end users (the MS moving between the two cells and its correspondent node or server) must implement the MSCTP protocol.

The architecture of simulated model proposed for the interconnection between IEEE 802.16e and IEEE 802.11s using MSCTP is illustrated in the figure 5 below:

Fig. 5. Interconnection model using MSCTP

The exchange of information between the MS and its correspondent node (CN) during the MS mobility scenario between IEEE 802.16e and IEEE 802.11s using MSCTP is illustrated in the figure 6 below:

Fig. 6. Messages exchanged during simulation of the protocol MSCTP

Interaction and Interconnection Between 802.16e & 802.11s 305

For the simulations, we choose to use Proxy Mobile IP (PMIP) [16] as layer 3 protocol that

PMIP is an amelioration of MIP, it introduces a functional entity called Proxy Mobile IP to help MIP traversal across VPN or "NAT and VPN" gateways. The PMIP is in the path between MS and its corresponding HA (Home Agent), and acts as a surrogate MS and HA. PMIP does not involve a change in the point of attachment address when the user moves,

During the simulations and with the two handover mechanisms, we will test three traffics types: the VoIP with a fixed size to 160 byte and a rate of 300 packet/sec, the Data with a fixed size to 640 bytes and a rate of 200 packet/sec, and the video streaming with a fixed size to 1280 bytes and a rate of 100 packet/sec (optimal values usually chosen in NS2).

Under NS2, with MSCTP protocol we use the CBR over MSCTP traffic type; and with IEEE

The duration of one simulation are fixed to 250 seconds, and the results are calculate every 10 seconds in 802.16e or 802.11s cell area, and every 5 seconds during the handover process

For the two networks: IEEE 802.16s and IEEE 802.11s, the simulation parameters under NS2

Two-Ray Ground [17]

*t t rt r <sup>r</sup> <sup>4</sup> PGG h h P (d) d L*

*2 2*

Transmission Power (Pt\_) 15 W 0.2818 W

Coverage Radius (Distance D) 1 km 300 m

Modulation OFDMA OFDM Frequency (Freq\_) 3.5 Ghz 2.4 Ghz

Receiving Threshold (RXThresh\_) 7.59375e-11 W 1.76148e-10 W

(CSThresh\_) 4.34219e-12 W 3.32874e-11 W

IEEE 802.16e IEEE 802.11s

and there is no need for the terminal to implement a client MIP stack.

interacts with lower layers via MIH module.

**3.4 Simulation parameters under NS2 simulator** 

802.21 architecture, we use the CBR over UDP traffic type.

in common area.

are illustrated in the table 2 below:

Carrier Sending Threshold

Radio Propagation Model

Table 2. Simulation parameters

Transmit Antenna Gain (Gt\_) 1 dB Receive Antenna Gain (Gr\_) 1 dB System Loss (L\_) 1 dB Transmit Antenna Height (ht\_) 1.5 m Receive Antenna Height (hr\_) 1.5 m

#### **3.3 Simulation model based on 802.21 architecture**

Based on the common model already described; with IEEE 802.21 architecture, an MIH server must be implemented in the Internet network, more precisely between the WiMAX CSN, and the WIFI CSN; and MIH modules must be implemented in the two CSNs and in the MS.

The architecture of the interconnection model proposed between the two networks using IEEE 802.21 is illustrated in the figure 7 below:

Fig. 7. Interconnection model using IEEE 802.21

The exchange of information between the MS and its correspondent node (CN) during the MS mobility scenario between IEEE 802.16e and IEEE 802.11s using the MIH module is illustrated in the figure 8 below:

Fig. 8. Messages exchanged during simulation of the protocol MIH

Based on the common model already described; with IEEE 802.21 architecture, an MIH server must be implemented in the Internet network, more precisely between the WiMAX CSN, and the WIFI CSN; and MIH modules must be implemented in the two CSNs and in the MS.

The architecture of the interconnection model proposed between the two networks using

The exchange of information between the MS and its correspondent node (CN) during the MS mobility scenario between IEEE 802.16e and IEEE 802.11s using the MIH module is

**3.3 Simulation model based on 802.21 architecture** 

IEEE 802.21 is illustrated in the figure 7 below:

Fig. 7. Interconnection model using IEEE 802.21

Fig. 8. Messages exchanged during simulation of the protocol MIH

illustrated in the figure 8 below:

For the simulations, we choose to use Proxy Mobile IP (PMIP) [16] as layer 3 protocol that interacts with lower layers via MIH module.

PMIP is an amelioration of MIP, it introduces a functional entity called Proxy Mobile IP to help MIP traversal across VPN or "NAT and VPN" gateways. The PMIP is in the path between MS and its corresponding HA (Home Agent), and acts as a surrogate MS and HA.

PMIP does not involve a change in the point of attachment address when the user moves, and there is no need for the terminal to implement a client MIP stack.

#### **3.4 Simulation parameters under NS2 simulator**

During the simulations and with the two handover mechanisms, we will test three traffics types: the VoIP with a fixed size to 160 byte and a rate of 300 packet/sec, the Data with a fixed size to 640 bytes and a rate of 200 packet/sec, and the video streaming with a fixed size to 1280 bytes and a rate of 100 packet/sec (optimal values usually chosen in NS2).

Under NS2, with MSCTP protocol we use the CBR over MSCTP traffic type; and with IEEE 802.21 architecture, we use the CBR over UDP traffic type.

The duration of one simulation are fixed to 250 seconds, and the results are calculate every 10 seconds in 802.16e or 802.11s cell area, and every 5 seconds during the handover process in common area.


For the two networks: IEEE 802.16s and IEEE 802.11s, the simulation parameters under NS2 are illustrated in the table 2 below:

Table 2. Simulation parameters

Interaction and Interconnection Between 802.16e & 802.11s 307

With the two speeds: 18 and 36 km/h, the QoS level is accepted for the VoIP traffic that needed a minimum delay of 100 ms [18]; but with the speed equal to 72 km/h, the QoS

Finally, with a speed of 18 km/h, the QoS is better in 802.11s network; with 36 km/h the QoS is equivalent in the two cells; and with 72 km/h of speed the QoS is better in 802.16e cell.

Fig. 10. Delay of HO from 802.16e to 802.11s / VoIP

Fig. 11. Delay of HO from 802.11s to 802.16e / Data

We present now the results of data traffic (see figures 11 and 12):

degrades during the handover.

#### **3.5 Performance criteria**

The performance criteria adopted in our simulations to compare MSCTP and IEEE 802.21 mechanisms in the case of vertical handover between 802.11s and 802.16e networks are: End-to-end delay, packets loss ratio and debit. These parameters are the main criteria of QoS measuring in the networks. To evaluate the QoS degree of these criteria, we will compare simulation results obtained with theoretical thresholds estimated to evaluate the QoS depending on traffic type.

#### **4. Results**

#### **4.1 End-to-end delays**

The end-to-end delay is a very important parameter to evaluate the QoS for the real time traffic. It is the time needed for a packet to be transmitted across a network from source to destination.

In this section, we will calculate the delays of packets during the simulation time for the three mobile speeds and the three traffic types; with the two vertical handover techniques: MIH architecture and MSCTP protocol; and applying the two scenarios: handover from 802.16e to 802.11s and from 802.11s to 802.16e.

We start by present the results for the VoIP traffic (see figures 9 and 10).

In the two figures 9 and 10, first for all the curves we see that during the handover process, the delays obtained with MSCTP are slightly lower than those obtained with the MIH; and the handover are executed with MIH before that with MSCTP.

Fig. 9. Delay of HO from 802.11s to 802.16e / VoIP

The performance criteria adopted in our simulations to compare MSCTP and IEEE 802.21 mechanisms in the case of vertical handover between 802.11s and 802.16e networks are: End-to-end delay, packets loss ratio and debit. These parameters are the main criteria of QoS measuring in the networks. To evaluate the QoS degree of these criteria, we will compare simulation results obtained with theoretical thresholds estimated to evaluate the QoS

The end-to-end delay is a very important parameter to evaluate the QoS for the real time traffic. It is the time needed for a packet to be transmitted across a network from source to

In this section, we will calculate the delays of packets during the simulation time for the three mobile speeds and the three traffic types; with the two vertical handover techniques: MIH architecture and MSCTP protocol; and applying the two scenarios: handover from

In the two figures 9 and 10, first for all the curves we see that during the handover process, the delays obtained with MSCTP are slightly lower than those obtained with the MIH; and

We start by present the results for the VoIP traffic (see figures 9 and 10).

the handover are executed with MIH before that with MSCTP.

**3.5 Performance criteria** 

depending on traffic type.

**4.1 End-to-end delays** 

802.16e to 802.11s and from 802.11s to 802.16e.

Fig. 9. Delay of HO from 802.11s to 802.16e / VoIP

**4. Results** 

destination.

Fig. 10. Delay of HO from 802.16e to 802.11s / VoIP

With the two speeds: 18 and 36 km/h, the QoS level is accepted for the VoIP traffic that needed a minimum delay of 100 ms [18]; but with the speed equal to 72 km/h, the QoS degrades during the handover.

Finally, with a speed of 18 km/h, the QoS is better in 802.11s network; with 36 km/h the QoS is equivalent in the two cells; and with 72 km/h of speed the QoS is better in 802.16e cell.

We present now the results of data traffic (see figures 11 and 12):

Fig. 11. Delay of HO from 802.11s to 802.16e / Data

Interaction and Interconnection Between 802.16e & 802.11s 309

With the video traffic, the delay values increase comparing by the VoIP or the data traffic. With the speed of 18 km/h, the delays not exceed 100 ms; with the speed of 36 km/h the delays exceed slightly 100 ms during the handover; and with the speed of 72 km/h, the

We calculate in this section the percentage of lost packets with the same cases as those

delay values exceed largely 100 ms during the handover and in the 802.11s cell.

described in the section 4.1. We start with the VoIP traffic (see figures 15 and 16):

Fig. 15. Percentage of packets lost with HO from 802.11s to 802.16e / VoIP

The delays with MSCTP are slightly lower than those with MIH.

Fig. 14. Delay of HO from 802.16e to 802.11s / Video

**4.2 Packet loss ratio** 

Fig. 12. Delay of HO from 802.16e to 802.11s / Data

With the data traffic, there is no delay constraint, so we can consider the QoS level acceptable for all the cases.

We note that the delays obtained with data are higher than those obtained with the VoIP traffic; for example with MSCTP protocol and with a speed of 18 km/h, the maximum delay obtained of the VoIP traffic is 65 ms versus 73 ms of the data traffic.

We present finally in this section the results of video streaming traffic (see figures 13 & 14):

Fig. 13. Delay of HO from 802.11s to 802.16e / Video

Fig. 14. Delay of HO from 802.16e to 802.11s / Video

With the video traffic, the delay values increase comparing by the VoIP or the data traffic.

With the speed of 18 km/h, the delays not exceed 100 ms; with the speed of 36 km/h the delays exceed slightly 100 ms during the handover; and with the speed of 72 km/h, the delay values exceed largely 100 ms during the handover and in the 802.11s cell.

The delays with MSCTP are slightly lower than those with MIH.

#### **4.2 Packet loss ratio**

308 Advanced Transmission Techniques in WiMAX

With the data traffic, there is no delay constraint, so we can consider the QoS level

We note that the delays obtained with data are higher than those obtained with the VoIP traffic; for example with MSCTP protocol and with a speed of 18 km/h, the maximum delay

We present finally in this section the results of video streaming traffic (see figures 13 & 14):

Fig. 12. Delay of HO from 802.16e to 802.11s / Data

Fig. 13. Delay of HO from 802.11s to 802.16e / Video

obtained of the VoIP traffic is 65 ms versus 73 ms of the data traffic.

acceptable for all the cases.

We calculate in this section the percentage of lost packets with the same cases as those described in the section 4.1. We start with the VoIP traffic (see figures 15 and 16):

Fig. 15. Percentage of packets lost with HO from 802.11s to 802.16e / VoIP

Interaction and Interconnection Between 802.16e & 802.11s 311

Fig. 18. Percentage of packets lost with HO from 802.16e to 802.11s / Data

Fig. 19. Percentage of packets lost with HO from 802.11s to 802.16e / Video

We pass now to video streaming traffic (see figures 19 and 20):

MIH during the Handover.

and in the 802.11s cell.

With the data traffic, the percentage of loss is zero with a speed of 18 km/h, and it is acceptable with a speed of 36 km/h. It has a maximum of 0.1% with MSCTP and 0.2% with

But with a speed of 72 km/h, the percentage of loss is not acceptable during the handover

Fig. 16. Percentage of packets lost with HO from 802.16e to 802.11s / VoIP

During the Handover, the only results that converge to the threshold of 1% [18] required by the VoIP traffic] are those corresponding to the speed of 72 km/h. For the two other speeds the QoS level is acceptable.

We present now the results of data traffic (see figures 17 and 18):

Fig. 17. Percentage of packets lost with HO from 802.11s to 802.11e / Data

Fig. 16. Percentage of packets lost with HO from 802.16e to 802.11s / VoIP

Fig. 17. Percentage of packets lost with HO from 802.11s to 802.11e / Data

We present now the results of data traffic (see figures 17 and 18):

the QoS level is acceptable.

During the Handover, the only results that converge to the threshold of 1% [18] required by the VoIP traffic] are those corresponding to the speed of 72 km/h. For the two other speeds

Fig. 18. Percentage of packets lost with HO from 802.16e to 802.11s / Data

With the data traffic, the percentage of loss is zero with a speed of 18 km/h, and it is acceptable with a speed of 36 km/h. It has a maximum of 0.1% with MSCTP and 0.2% with MIH during the Handover.

But with a speed of 72 km/h, the percentage of loss is not acceptable during the handover and in the 802.11s cell.

We pass now to video streaming traffic (see figures 19 and 20):

Fig. 19. Percentage of packets lost with HO from 802.11s to 802.16e / Video

Interaction and Interconnection Between 802.16e & 802.11s 313

Fig. 21. Debit of HO from 802.11s to 802.16e / VoIP

Fig. 22. Debit of HO from 802.16e to 802.11s / VoIP

We present now the results of data traffic (see figures 23 and 24):

Fig. 20. Percentage of packets lost with HO from 802.16e to 802.11s / Video

With the video traffic, the loss values are higher than those with VoIP or data traffic, and with a speed of 18 km/h the results are acceptable because they not exceed 1% which is a maximum of loss required for video traffic [18].

We note that the case of handover from 802.16e to 802.11s produce results slightly better than those obtained with the opposite case during the handover.

With the speed of 36 km/h, the maximum of the % of loss exceeds slightly 1% with MIH, and is equal to 1% with MSCTP; and the results are equivalent in the two ways of handover.

With a speed of 72 km/h, the results are not acceptable with the two mechanisms during the handover because they exceed largely 1%, and the results of the handover from 802.11s to 802.16e are better than those of the opposite case. And finally with this speed, the results are acceptable in the 802.16e cell but not in the 802.11s cell.

#### **4.3 Debit**

Finally, in this section we will evaluate the debit experimented by the mobile during the handover cases already proposed for the simulations in section 4.1. We start with the VoIP traffic (see figures 21 and 22).

Concerning the debit of VoIP traffic and with a minimum required by this type of traffic fixed to 4 kb/s [18]; the all cases proposed with the two speeds: 18 and 36 km/h present an acceptable debit; but with the speed of 72 km/h and during the handover, the results are lower than 4 kb/s, and the debit obtained with MSCTP during the handover is slightly higher than that obtained with MIH.

Fig. 20. Percentage of packets lost with HO from 802.16e to 802.11s / Video

than those obtained with the opposite case during the handover.

acceptable in the 802.16e cell but not in the 802.11s cell.

maximum of loss required for video traffic [18].

handover.

**4.3 Debit** 

traffic (see figures 21 and 22).

higher than that obtained with MIH.

With the video traffic, the loss values are higher than those with VoIP or data traffic, and with a speed of 18 km/h the results are acceptable because they not exceed 1% which is a

We note that the case of handover from 802.16e to 802.11s produce results slightly better

With the speed of 36 km/h, the maximum of the % of loss exceeds slightly 1% with MIH, and is equal to 1% with MSCTP; and the results are equivalent in the two ways of

With a speed of 72 km/h, the results are not acceptable with the two mechanisms during the handover because they exceed largely 1%, and the results of the handover from 802.11s to 802.16e are better than those of the opposite case. And finally with this speed, the results are

Finally, in this section we will evaluate the debit experimented by the mobile during the handover cases already proposed for the simulations in section 4.1. We start with the VoIP

Concerning the debit of VoIP traffic and with a minimum required by this type of traffic fixed to 4 kb/s [18]; the all cases proposed with the two speeds: 18 and 36 km/h present an acceptable debit; but with the speed of 72 km/h and during the handover, the results are lower than 4 kb/s, and the debit obtained with MSCTP during the handover is slightly

Fig. 21. Debit of HO from 802.11s to 802.16e / VoIP

Fig. 22. Debit of HO from 802.16e to 802.11s / VoIP

We present now the results of data traffic (see figures 23 and 24):

Interaction and Interconnection Between 802.16e & 802.11s 315

With the data traffic, the debit decrease when the speed increase. The debit is better in 802.11s cell when the speed is weak, and it is better in 802.16e when the speed is high.

We present finally the results of video traffic (see figures 25 and 26):

Fig. 25. Debit of HO from 802.11s to 802.16e / Video

Fig. 23. Debit of HO from 802.11s to 802.16e / Data

Fig. 24. Debit of HO from 802.16e to 802.11s / Data

Fig. 23. Debit of HO from 802.11s to 802.16e / Data

Fig. 24. Debit of HO from 802.16e to 802.11s / Data

With the data traffic, the debit decrease when the speed increase. The debit is better in 802.11s cell when the speed is weak, and it is better in 802.16e when the speed is high.

We present finally the results of video traffic (see figures 25 and 26):

Fig. 25. Debit of HO from 802.11s to 802.16e / Video

Interaction and Interconnection Between 802.16e & 802.11s 317

Also the handover from 802.11s to 802.16e generates results better than the opposite case of handover. But with a high speed, it is the opposite rather because the mobile WIMAX supports better the increasing speeds; and also the results in this case are still not acceptable

It should be noted that during all the simulations, the scenarios proposed does not include

For future work, we will propose interconnection models between networks of different family, we will mix a network world with a telecommunication world, and we will try to

[1] The Information Science Institute (ISI), *"The Network Simulator-NS-2"*,

[2] EEE Std, "Air Interface for Fixed and Mobile Broadband Wireless Access Systems," IEEE

[3] IEEE Std, "Air Interface for Fixed Broadband Wireless Access Systems," Local and

[5] WiMAX Forum, *"WiMAX End-to-End Network Systems Architecture,"* Draft Stage 2: Architecture Tenets, Reference Model and Reference Points, June 2007. [6] Steven Conner, Jan Kruys, Kyeongsoo Kim and Juan Carlos Zuniga, *"IEEE 802.11s* 

[7] Guido R. Hiertz, Sebastian Max, Rui Zhao, Dee Denteneer and Lars Berlemann,

[9] Stewart R., & al., IETF, *"Stream Control Transmission Protocol (SCTP) Dynamic Address* 

[10] Koh, S., & al., *"mSCTP for Soft Handover in Transport Layer,"* IEEE Communication

[11] Memory graduation, Esteban Zimanyi, "Performance analysis of vertical Handover

[12] Deng Feng, "Seamless Handover between CDMA2000 and 802.11 WLAN using

[14] Jared Stein, "Survey of IEEE 802.21 Media Independent Handover Services," April 2006. [15] V. Gupta, "IEEE 802.21 standard and metropolitan area networks: Media Independent

[16] K. Leung, G. Dommety, P. Yegani & K. Chowdhury, *"Mobility Management Using Proxy* 

*Tutorial,"* Overview of the Amendment for Wireless Local Area Mesh Networking,

*"Principles of IEEE 802.11s,"* Computer Communications and Networks, 2007,

*Reconfiguration,"* IETF Internet, Draft, draft-ietf-tsvwg-addip-sctp-13.txt, November

propose a handover mechanism adapted to the two entities that we will define.

[4] arviz Yegani, "*WiMAX Overview,*" White paper, IETF-64 Cisco Systems, 2005.

comparing by QoS level needed for each traffic type.

http://www.isi.edu/ nsnam/ns/.

IEEE 802 Planary, November 2006.

ICCCN 2007.

mSCTP," Thesis, 2006. [13] IEEE 802.21 tutorial, July 2006.

2005.

metropolitan area networks, Part 16, 2004.

[8] RFC 2960, "Stream Control Transmission Protocol," IETF, 2000.

Letters, Vol. 8, No.3, pp.189-191, March 2004.

between UMTS and 802.11 networks," 2005.

*Mobile IPv4"*, Internet Draft, IETF, 2007.

Handover services", Draft P802.21/D05.00, April 2007.

802.16e, Part 16, February 2006.

cell congestion or lack of available resources.

**6. References** 

Fig. 26. Debit of HO from 802.16e to 802.11s / Video

With the video traffic, the debit values decrease comparing to the other traffic types.

#### **5. Conclusion**

The interoperability and the vertical handover between different networks present currently a real challenge to overcome. The difference of networks operation is the main reason of this problem. And, for pass to the 4G networks, it is important to resolve this problem of interoperability between different networks.

Our work has focused on the interconnection between two wireless radio networks of the IEEE 802 family, and we are concentrating on the QoS aspect for several traffics types especially during the handover process. For doing that, we have proposed two interconnection models based on two recent handover mechanisms, and we have simulated those two models with three mobile speeds and in the both directions of networks.

Observing the results obtained, we can conclude that with a low or medium speed of displacement of a mobile station, the both techniques: IEEE 802.21 and MSCTP present a good solution during the vertical handover. With the two techniques, there are very few interruptions during the vertical handover. But based on details of simulation results, we notice that with MSCTP protocol we obtained a QoS level slightly better than that obtained with MIH architecture.

Also the handover from 802.11s to 802.16e generates results better than the opposite case of handover. But with a high speed, it is the opposite rather because the mobile WIMAX supports better the increasing speeds; and also the results in this case are still not acceptable comparing by QoS level needed for each traffic type.

It should be noted that during all the simulations, the scenarios proposed does not include cell congestion or lack of available resources.

For future work, we will propose interconnection models between networks of different family, we will mix a network world with a telecommunication world, and we will try to propose a handover mechanism adapted to the two entities that we will define.

#### **6. References**

316 Advanced Transmission Techniques in WiMAX

Fig. 26. Debit of HO from 802.16e to 802.11s / Video

interoperability between different networks.

**5. Conclusion** 

with MIH architecture.

With the video traffic, the debit values decrease comparing to the other traffic types.

The interoperability and the vertical handover between different networks present currently a real challenge to overcome. The difference of networks operation is the main reason of this problem. And, for pass to the 4G networks, it is important to resolve this problem of

Our work has focused on the interconnection between two wireless radio networks of the IEEE 802 family, and we are concentrating on the QoS aspect for several traffics types especially during the handover process. For doing that, we have proposed two interconnection models based on two recent handover mechanisms, and we have simulated

Observing the results obtained, we can conclude that with a low or medium speed of displacement of a mobile station, the both techniques: IEEE 802.21 and MSCTP present a good solution during the vertical handover. With the two techniques, there are very few interruptions during the vertical handover. But based on details of simulation results, we notice that with MSCTP protocol we obtained a QoS level slightly better than that obtained

those two models with three mobile speeds and in the both directions of networks.


**15** 

*Malaysia* 

**Inter-Domain Handover in WiMAX Networks** 

The most attractive feature of WiMAX is arguable the mobility capability that IEEE 802.16e (IEEE, 2004) standard adds to the previous standard. With mobility support, handover has become one of the most important factors that impact the performance of IEEE 802.16e system. Handover is the process of maintaining active sessions of a mobile station when it migrates from current base station to target base station area. Handover occurs when a mobile station changes its point of attachment on the network. However during hard handover, the mobile station cannot receive or send any packet for a short time interval. This is referred to system disruption time because the services are interrupted or handover latency. In WiMAX, when a mobile node or mobile station changes its location, it moves the

 The mobile station changes its point of attachment between the base stations which reside in the same Access Services Network (ASN) that is called ASN-anchored, intra, micro, or layer 2 handover. In an ASN-anchored handover, the mobile station resides within previous network address (both current and target base stations located in the same IP subnet). In this scenario, the mobile station does not change its IP

 The mobile station or mobile node changes its point of attachment between the base stations which reside in different ASN (different IP subnets) that is called Connectivity Services Network (CSN)-anchored, inter, macro, or layer 3 handover. In a CSNanchored handover, in addition to link layer handover a mobile node must perform a

The intra-domain handover procedure requires support from the physical and MAC layers. IEEE 802.16e has its own MAC layer or layer 2 handover algorithm, but a layer 3 handover algorithm is also required to support the Internet Protocol (IP) addressing, for inter-domain handover. A typical protocol in network layer for mobile terminals is Mobile IP include Mobile IPv4 (MIPv4), (IEEE, 2002) and Mobile IPv6 (MIPv6), (IEEE, 2004) that have been standardized by the Internet Engineering Task Force (IETF). There are many problems associated with MIPv4, such as triangular routing, security and limitation of address space which were solved by using MIPv6. But there still remain some other problems, such as long

service disruption time (handover latency), signalling overhead and packet loss.

point of attachment to the network in two different scenarios;

configuration, only link layer is re-established.

new IP configuration to avoid disconnection.

**1. Introduction** 

**Using Optimized Fast Mobile IPv6** 

Seyyed Masoud Seyyedoshohadaei,

*Universiti Putra Malaysia (UPM),* 

Borhanuddin Mohd Ali and Sabira Khatun


### **Inter-Domain Handover in WiMAX Networks Using Optimized Fast Mobile IPv6**

Seyyed Masoud Seyyedoshohadaei, Borhanuddin Mohd Ali and Sabira Khatun *Universiti Putra Malaysia (UPM), Malaysia* 

#### **1. Introduction**

318 Advanced Transmission Techniques in WiMAX

[17] Information Sciences Institute (ISI), *"NSNAM web pages, 18.2 Two-Ray Ground reflection model,"* http://www.isi.edu/nsnam/ns/doc/node218.html, January 2009. [18] WiMAX Community, "WiMAX fundamentals, 1.7.3 Quality of Service", June 2007.

> The most attractive feature of WiMAX is arguable the mobility capability that IEEE 802.16e (IEEE, 2004) standard adds to the previous standard. With mobility support, handover has become one of the most important factors that impact the performance of IEEE 802.16e system. Handover is the process of maintaining active sessions of a mobile station when it migrates from current base station to target base station area. Handover occurs when a mobile station changes its point of attachment on the network. However during hard handover, the mobile station cannot receive or send any packet for a short time interval. This is referred to system disruption time because the services are interrupted or handover latency. In WiMAX, when a mobile node or mobile station changes its location, it moves the point of attachment to the network in two different scenarios;


The intra-domain handover procedure requires support from the physical and MAC layers. IEEE 802.16e has its own MAC layer or layer 2 handover algorithm, but a layer 3 handover algorithm is also required to support the Internet Protocol (IP) addressing, for inter-domain handover. A typical protocol in network layer for mobile terminals is Mobile IP include Mobile IPv4 (MIPv4), (IEEE, 2002) and Mobile IPv6 (MIPv6), (IEEE, 2004) that have been standardized by the Internet Engineering Task Force (IETF). There are many problems associated with MIPv4, such as triangular routing, security and limitation of address space which were solved by using MIPv6. But there still remain some other problems, such as long service disruption time (handover latency), signalling overhead and packet loss.

Inter-Domain Handover in WiMAX Networks Using Optimized Fast Mobile IPv6 321

In MIPv6 also, each mobile node has two addresses, a static home address under its home network (HoA), and a care of address (CoA) as the mobile node roams to a foreign network for packet routing. The mobile node can create a CoA from a router advertisement message sent by the new visited network. When the mobile node moves to a foreign network, the mobile node sends Binding Update (BU) messages with its CoA to the home agent in order to update the home agent of its current point of attachment. In this way, mobile node's home agent can always detect coming communication packets to mobile node with home address of mobile node, and dispatching these packets to the mobile nodes' CoA via dynamically created IP tunnels. The signalling and data traffic are all transmitted via a unified IP framework, because, all the MIPv6 signalling messages are formed by extending IP protocols with option headers. However the MIPv6 causes a long latency problem. In order to improve handover performance of MIPv6, IETF introduced some IPv6 mobility

In MIPv6, the movement detection (based on Router Advertisement in IP-layer) and the address configuration procedures cause a long latency problem. FMIPv6 decreases delay of the movement detection and the address configuration phases of MIPv6. It enables the mobile node to provide the target base station identifier (BSID) and detects upcoming entrance to new subnet. It therefore reduces delay of movement detection. For new address configuration, in the FMIPv6 the mobile node obtains the new associated subnet prefix

After the mobile node select one of the candidate base stations as target base station according to its policy, it sends the Router Solicitation for Proxy (RtSolPr) to the current access router or previous access router and receives Proxy Router Advertisement (PrRtAdv) messages in return. During exchanges of these messages the mobile node obtains the subnet prefix of the target base station. The current base station configures a new IP address (CoA) based on the subnet prefix of the target base station. After that, the mobile node sends a Fast Binding Update (FBU) message to the previous access router. The purpose of FBU messages is to inform the access router that there is a binding between the current CoA at the current subnet and the new CoA (NCoA) at the target subnet. Then, the Handover Initiation (HI) message is sent to the target or new access router by previous access router. The new access router performs duplicate address detection (DAD) to check validity of NCoA. After DAD procedure the new access router reply with handover acknowledge (HAck) message to the current access router. At this instant, a tunnel between the CoA of and NCoA of mobile node is established. The previous access router sends a fast-binding acknowledgement (FBAck) message to new access router. Fig. 1 illustrates the FMIPv6 procedure for Predictive and Reactive mode. If the mobile node receives the FBAck message in the current subnet before the layer 2 handover is started (there is enough time to exchange required messages to establish tunnel), handover occurs in the predictive mode. Otherwise, if the mobile node is forced to move to the new access router without receiving FBAck, FMIPv6 is in reactive

In the predictive mode, the previous access router first store the tunnelled packets in a buffer. After the mobile node attaches to the new link, mobile node sends a Fast Neighbour

information in advance, while it is still connecting to the current subnet.

protocol solutions such as HMIPv6 and FMIPv6.

**2.1.2 Fast mobile IPv6** 

mode.

However, MIPv6 does not solve the handover latency problem which is not negligible for real-time applications such as video streaming and Voice over IP (VoIP). Proxy Mobile IPv6 (PMIPv6), Hierarchical Mobile IPv6 (HMIPv6) and Fast Mobile IPv6 (FMIPv6) have been proposed to decrease long handover latency of MIPv6. The MIPv6 Signalling and Handoff Optimization (MIPSHOP) working group has standardized FMIPv6 (IETF, 2005). FMIPv6 is capable decreasing the handover latency and packet loss by mobility detection and creating new address for the target network and receives data through tunnelling in advance. Because of this, FMIPv6 is used as IP layer protocol in WiMAX. However, due to complexity of handover pattern, designing an impressive handover process to support all mobility scenarios with acceptable latency is still a challenge. There have been many proposals on how to effectively coordinate the FMIPv6 handover algorithm in layer 3 with handover algorithm of the IEEE 802.16e system in layer 2. To overcome some of the shortcomings in the proposed proposals an Optimized Fast Handover Scheme (OFHS) is proposed and presented in this chapter.

This chapter is organized as follows. In section 2, the MIPv6, FMIPv6, IEEE 802.16e handover and related works are described. The proposed scheme is explained in section 3. In section 4, a numerical model is developed to evaluate the performance of OFHS compared with that of RFC5270 (IEEE, 2008). T The results and discussion are presented in section 5, and finally, in section 6, conclusions of this chapter are made.

#### **2. Background and previous works**

In this section first, some literature that needed to explain proposed method such as mobile IP and the layer 2 handover procedures in IEEE 802.16.e or mobile WiMAX are described. Then some related works are introduced which have focused on how apply FMIPv6 over IEEE 802.16e to support inter-domain handover.

#### **2.1 Background**

When a host moved to other subnet, the IP address became incorrect for routing and if hosts used new IP address the connections would be terminated because the new IP address was unknown. Mobile IP mechanism works based on a temporary IP address named Care of Address (CoA). The MIPv4 and MIPv6 have introduced for difference IP addressing. In this work IPv6 has been used for addressing. Therefore, in following sections (2.1.1 and 2.1.2) MIPv6 and Fast MIPv6 are described.

The IEEE 802.16e standard supports mobile user in WiMAX network. It supports only intradomain handover that movement of the mobile station with in same subnet does not affect the IP address. In section 2.1.3, layer 2 handover procedure that has been defined in IEEE 802.16e explined.

#### **2.1.1 Mobile IPv6**

The MIPv6 is a protocol to support inter-domain mobility (in network layer) for IPv6 based network. In MIPv6, the packets that are sent to the mobile node from the correspondent node are intercepted and forwarded by a home agent. The MIPv6 has same functions as MIPv4 that is adapted for MIPv6.

In MIPv6 also, each mobile node has two addresses, a static home address under its home network (HoA), and a care of address (CoA) as the mobile node roams to a foreign network for packet routing. The mobile node can create a CoA from a router advertisement message sent by the new visited network. When the mobile node moves to a foreign network, the mobile node sends Binding Update (BU) messages with its CoA to the home agent in order to update the home agent of its current point of attachment. In this way, mobile node's home agent can always detect coming communication packets to mobile node with home address of mobile node, and dispatching these packets to the mobile nodes' CoA via dynamically created IP tunnels. The signalling and data traffic are all transmitted via a unified IP framework, because, all the MIPv6 signalling messages are formed by extending IP protocols with option headers. However the MIPv6 causes a long latency problem. In order to improve handover performance of MIPv6, IETF introduced some IPv6 mobility protocol solutions such as HMIPv6 and FMIPv6.

#### **2.1.2 Fast mobile IPv6**

320 Advanced Transmission Techniques in WiMAX

However, MIPv6 does not solve the handover latency problem which is not negligible for real-time applications such as video streaming and Voice over IP (VoIP). Proxy Mobile IPv6 (PMIPv6), Hierarchical Mobile IPv6 (HMIPv6) and Fast Mobile IPv6 (FMIPv6) have been proposed to decrease long handover latency of MIPv6. The MIPv6 Signalling and Handoff Optimization (MIPSHOP) working group has standardized FMIPv6 (IETF, 2005). FMIPv6 is capable decreasing the handover latency and packet loss by mobility detection and creating new address for the target network and receives data through tunnelling in advance. Because of this, FMIPv6 is used as IP layer protocol in WiMAX. However, due to complexity of handover pattern, designing an impressive handover process to support all mobility scenarios with acceptable latency is still a challenge. There have been many proposals on how to effectively coordinate the FMIPv6 handover algorithm in layer 3 with handover algorithm of the IEEE 802.16e system in layer 2. To overcome some of the shortcomings in the proposed proposals an Optimized Fast Handover Scheme (OFHS) is proposed and

This chapter is organized as follows. In section 2, the MIPv6, FMIPv6, IEEE 802.16e handover and related works are described. The proposed scheme is explained in section 3. In section 4, a numerical model is developed to evaluate the performance of OFHS compared with that of RFC5270 (IEEE, 2008). T The results and discussion are presented in

In this section first, some literature that needed to explain proposed method such as mobile IP and the layer 2 handover procedures in IEEE 802.16.e or mobile WiMAX are described. Then some related works are introduced which have focused on how apply FMIPv6 over

When a host moved to other subnet, the IP address became incorrect for routing and if hosts used new IP address the connections would be terminated because the new IP address was unknown. Mobile IP mechanism works based on a temporary IP address named Care of Address (CoA). The MIPv4 and MIPv6 have introduced for difference IP addressing. In this work IPv6 has been used for addressing. Therefore, in following sections (2.1.1 and 2.1.2)

The IEEE 802.16e standard supports mobile user in WiMAX network. It supports only intradomain handover that movement of the mobile station with in same subnet does not affect the IP address. In section 2.1.3, layer 2 handover procedure that has been defined in IEEE

The MIPv6 is a protocol to support inter-domain mobility (in network layer) for IPv6 based network. In MIPv6, the packets that are sent to the mobile node from the correspondent node are intercepted and forwarded by a home agent. The MIPv6 has same functions as

section 5, and finally, in section 6, conclusions of this chapter are made.

presented in this chapter.

**2.1 Background** 

802.16e explined.

**2.1.1 Mobile IPv6** 

**2. Background and previous works** 

MIPv6 and Fast MIPv6 are described.

MIPv4 that is adapted for MIPv6.

IEEE 802.16e to support inter-domain handover.

In MIPv6, the movement detection (based on Router Advertisement in IP-layer) and the address configuration procedures cause a long latency problem. FMIPv6 decreases delay of the movement detection and the address configuration phases of MIPv6. It enables the mobile node to provide the target base station identifier (BSID) and detects upcoming entrance to new subnet. It therefore reduces delay of movement detection. For new address configuration, in the FMIPv6 the mobile node obtains the new associated subnet prefix information in advance, while it is still connecting to the current subnet.

After the mobile node select one of the candidate base stations as target base station according to its policy, it sends the Router Solicitation for Proxy (RtSolPr) to the current access router or previous access router and receives Proxy Router Advertisement (PrRtAdv) messages in return. During exchanges of these messages the mobile node obtains the subnet prefix of the target base station. The current base station configures a new IP address (CoA) based on the subnet prefix of the target base station. After that, the mobile node sends a Fast Binding Update (FBU) message to the previous access router. The purpose of FBU messages is to inform the access router that there is a binding between the current CoA at the current subnet and the new CoA (NCoA) at the target subnet. Then, the Handover Initiation (HI) message is sent to the target or new access router by previous access router. The new access router performs duplicate address detection (DAD) to check validity of NCoA. After DAD procedure the new access router reply with handover acknowledge (HAck) message to the current access router. At this instant, a tunnel between the CoA of and NCoA of mobile node is established. The previous access router sends a fast-binding acknowledgement (FBAck) message to new access router. Fig. 1 illustrates the FMIPv6 procedure for Predictive and Reactive mode. If the mobile node receives the FBAck message in the current subnet before the layer 2 handover is started (there is enough time to exchange required messages to establish tunnel), handover occurs in the predictive mode. Otherwise, if the mobile node is forced to move to the new access router without receiving FBAck, FMIPv6 is in reactive mode.

In the predictive mode, the previous access router first store the tunnelled packets in a buffer. After the mobile node attaches to the new link, mobile node sends a Fast Neighbour

Inter-Domain Handover in WiMAX Networks Using Optimized Fast Mobile IPv6 323

The reduction of inter-domain handover latency in IEEE 802.16e handover process had been presented in several papers. A link layer optimized scheme that reduces the link-layer handover latency by analyzing and optimizing each step of the procedure is suggested in (Lee, D. et al., 2006). In principle, the overall handover latency does not decrease by simple reduction of the link layer latency. To solve this problem, a cross-layer fast handover scheme for the IEEE 802.16e system is proposed in (Han et al., 2007). It coordinates FMIPv6 with IEEE 802.16e handover procedure to reduce the handover latency. This scheme with a little

FH802.16e is a cross layering design for FMIPv6 handover over IEEE 802.16e. One-way signaling is used in the majority of the existing cross layering handovers researches. They usually defined cross layer signals from MAC layer to IP layer. In the Han et al. scheme, two-way signaling between MAC layer and IP layer is defined. This concept helps to achieve faster handover algorithm than previous algorithms. For efficient handovers and reduce the handover latency the authors introduce one command and three events. Same events and command have been proposed in the IEEE 802.21 Media Independent Handover (MIH) (IETF, 2007). They support the interaction between both IP and MAC layers handover

NEW\_CANDIDATE\_BS\_FOUND: this includes the BSID(s) of candidate base station(s) and

is sent by MAC layer to IP layer (FMIPv6) when a new base station(s) is found.

Fig. 2. IEEE 802.16e handover procedure

change is used in RFC5270 (IETF, 2008).

procedures. The event are defined as follows:

**2.2 Related research works** 

**2.2.1 RFC 5270** 

Advertisement (FNA) message to the new access router. Upon reception of an FNA message, the new access router delivers the buffered packets to the mobile station.. In reactive mode, mobile node receives packets from the new access router after the packets are rerouted from previous to new access router.

Fig. 1. FMIPv6 Procedure Predictive mode and Reactive Mode

#### **2.1.3 IEEE 802.16e link layer handover**

The IEEE 802.16e layer 2 handover procedure can be divided into two steps: handover preparation and handover execution. Fig. 2 illustrates the IEEE 802.16e handover procedure.

The handover preparation can be initiated by either mobile station or base station. During this period, the neighbouring base stations are compared according to its policy. Some metrics such as Quality of Service (QoS) parameters or signal strength are considered to target base station selection. The current base station periodically sends the neighbour advertisement (MOB\_NBR-ADV) messages to mobile stations. This message contains information about neighbouring base stations, and the mobile station is capable to select target base stations for a future handover. In order to search for the suitability of neighbouring base stations, mobile station may execute a scanning operation (if necessary). It sends MOB\_SCN-REQ to current base station to obtain neighbouring base stations information and the base station reply by MOB\_SCN-RSP message. After a mobile station decides to perform handover, it sends a MOB\_MSHO-REQ message contain candidate base station identity to the current base station. The current base station negotiates with candidate base stations with exchanges HO-pre-notification and HO-pre-notificationresponse messages. Then the current base station introduces the recommended base stations by sending an MOB\_BSHO-RSP message to mobile station.

The handover execution is started by sending an MOB\_HO-IND message from mobile station to the current base station. This message contains selected target base station, and after that packet exchanging between mobile station and current base station is terminate. After IEEE 802.16e network entry process, the mobile station tuned its own parameters to the target base station. The buffered packets are sent to the mobile station from the target base station (it now becomes current base station). If the new base station has a new IP address, a network layer handover mechanism is needed.

Fig. 2. IEEE 802.16e handover procedure

#### **2.2 Related research works**

The reduction of inter-domain handover latency in IEEE 802.16e handover process had been presented in several papers. A link layer optimized scheme that reduces the link-layer handover latency by analyzing and optimizing each step of the procedure is suggested in (Lee, D. et al., 2006). In principle, the overall handover latency does not decrease by simple reduction of the link layer latency. To solve this problem, a cross-layer fast handover scheme for the IEEE 802.16e system is proposed in (Han et al., 2007). It coordinates FMIPv6 with IEEE 802.16e handover procedure to reduce the handover latency. This scheme with a little change is used in RFC5270 (IETF, 2008).

#### **2.2.1 RFC 5270**

322 Advanced Transmission Techniques in WiMAX

Advertisement (FNA) message to the new access router. Upon reception of an FNA message, the new access router delivers the buffered packets to the mobile station.. In reactive mode, mobile node receives packets from the new access router after the packets are

**PAR MOBILE** 

RtSol PrRtAd

FNA(FBU )

**NODE** 

The IEEE 802.16e layer 2 handover procedure can be divided into two steps: handover preparation and handover execution. Fig. 2 illustrates the IEEE 802.16e handover procedure. The handover preparation can be initiated by either mobile station or base station. During this period, the neighbouring base stations are compared according to its policy. Some metrics such as Quality of Service (QoS) parameters or signal strength are considered to target base station selection. The current base station periodically sends the neighbour advertisement (MOB\_NBR-ADV) messages to mobile stations. This message contains information about neighbouring base stations, and the mobile station is capable to select target base stations for a future handover. In order to search for the suitability of neighbouring base stations, mobile station may execute a scanning operation (if necessary). It sends MOB\_SCN-REQ to current base station to obtain neighbouring base stations information and the base station reply by MOB\_SCN-RSP message. After a mobile station decides to perform handover, it sends a MOB\_MSHO-REQ message contain candidate base station identity to the current base station. The current base station negotiates with candidate base stations with exchanges HO-pre-notification and HO-pre-notificationresponse messages. Then the current base station introduces the recommended base stations

FB

FBack

Delivered Packets

Forward Packets

**NA R**

The handover execution is started by sending an MOB\_HO-IND message from mobile station to the current base station. This message contains selected target base station, and after that packet exchanging between mobile station and current base station is terminate. After IEEE 802.16e network entry process, the mobile station tuned its own parameters to the target base station. The buffered packets are sent to the mobile station from the target base station (it now becomes current base station). If the new base station has a new IP

rerouted from previous to new access router.

**PAR NAR MOBILE** 

FBack FBack

Delivered Packets

**NODE** 

**2.1.3 IEEE 802.16e link layer handover** 

FNA

RtSolPr PrRtAd FBU

Fig. 1. FMIPv6 Procedure Predictive mode and Reactive Mode

HAck

Forward Packets

**DAD**

HI

by sending an MOB\_BSHO-RSP message to mobile station.

address, a network layer handover mechanism is needed.

FH802.16e is a cross layering design for FMIPv6 handover over IEEE 802.16e. One-way signaling is used in the majority of the existing cross layering handovers researches. They usually defined cross layer signals from MAC layer to IP layer. In the Han et al. scheme, two-way signaling between MAC layer and IP layer is defined. This concept helps to achieve faster handover algorithm than previous algorithms. For efficient handovers and reduce the handover latency the authors introduce one command and three events. Same events and command have been proposed in the IEEE 802.21 Media Independent Handover (MIH) (IETF, 2007). They support the interaction between both IP and MAC layers handover procedures. The event are defined as follows:

NEW\_CANDIDATE\_BS\_FOUND: this includes the BSID(s) of candidate base station(s) and is sent by MAC layer to IP layer (FMIPv6) when a new base station(s) is found.

Inter-Domain Handover in WiMAX Networks Using Optimized Fast Mobile IPv6 325

station in this message, a scanning may be performed to acquire more dynamic parameters for the new base stations. If the newly found base stations are candidates for the target BSs, the NLD event is delivered to its IP layer from the mobile node MAC layer with the found BSIDs. The Router Solicitation Proxy (RtSolPr) message and Proxy Router Advertisement (PrRtAdv) messages are exchanged between the mobile node and previous access router. The terminal initiates handover by sending a Mobile Handover Request (MOB\_MSHO-REQ) message to the current base station and receives a Mobile Handover Response (MOB\_BSHO-RSP) message in reply with a target base station in it. The current base station may also initiate handover by sending a MOB\_BSHO-REQ message to the mobile

After the mobile node receives MOB\_BSHO-REQ or MOB\_BSHO-RSP from the base station, the IP layer is triggered by link layer through a LHI to send Fast Binding Update (FBU) to the previous access router. The Handover Indication (HI) and Handover indication Acknowledge (Hack) messages are exchanged between previous and new access routers. The duplicate address detection is performed by new access router (it validates the uniqueness of NCoA in the new subnet, establishes tunnel and sends Fast Biding Acknowledge (FBack) message to the mobile station. Once the tunnel is established, the packets that are destined for the mobile node CoA are forwarded to the NCoA at the new access router through the tunnel. Upon receiving the FBack, the mobile node link layer is signalled by its network layer through a LSW to manage handover by sending a Mobile

node.

Fig. 4. FMIPv6 over IEEE 802.16e, Reactive Mode

LINK\_GOING\_DOWN: This is sent by MAC layer to IP layer (FMIPv6) when a mobile node receives an MOB\_BSHO-REQ or an MOB\_MSHO-RSP message which includes the target BSID. Upon receiving this event, the IP layer of the mobile node performs the handover preparation by sending an FBU message to the current access router.

LINK\_SWITCH: This is sent by IP layer (FMIPv6) to MAC layer when the IP layer of a mobile node receives an FBAck message. It caused the mobile node MAC layer start handover execution by sending an MOB\_HO-IND message to the current base station.

LINK\_UP: This is sent by MAC layer to IP layer (FMIPv6) to inform layer 3 that the network re-entry procedure of IEEE 802.16e is terminated. Upon receiving this event, the IP layer of mobile node sends an FNA message.

The scheme proposed in this article provides RFC5270 and the names of triggers change to: New Link Detected (NLD), Link Handover Impend (LHI), Link Switch (LSW), and Link Up (LUP). Fig. 3 and Fig. 4 show the message sequence diagram of the predictive and reactive FMIPv6 handover initiated by the RFC5270. The handover procedure of RFC5270 consists of two stages: handover preparation and handover execution. Just as FMIPv6 that supports all inter-domain handover scenarios, two modes (predictive and reactive) are defined in RFC5270.

Fig. 3. FMIPv6 over IEEE 802.16e, Predictive Mode

**Predictive Mode:** Here, the current base station generates and broadcasts a Mobile Neighbor Advertisement (MOB\_NBR-ADV) message periodically. It contains the network topology and static link layer information. When the mobile node discovers a new base

LINK\_GOING\_DOWN: This is sent by MAC layer to IP layer (FMIPv6) when a mobile node receives an MOB\_BSHO-REQ or an MOB\_MSHO-RSP message which includes the target BSID. Upon receiving this event, the IP layer of the mobile node performs the handover

LINK\_SWITCH: This is sent by IP layer (FMIPv6) to MAC layer when the IP layer of a mobile node receives an FBAck message. It caused the mobile node MAC layer start handover execution by sending an MOB\_HO-IND message to the current base station.

LINK\_UP: This is sent by MAC layer to IP layer (FMIPv6) to inform layer 3 that the network re-entry procedure of IEEE 802.16e is terminated. Upon receiving this event, the IP layer of

The scheme proposed in this article provides RFC5270 and the names of triggers change to: New Link Detected (NLD), Link Handover Impend (LHI), Link Switch (LSW), and Link Up (LUP). Fig. 3 and Fig. 4 show the message sequence diagram of the predictive and reactive FMIPv6 handover initiated by the RFC5270. The handover procedure of RFC5270 consists of two stages: handover preparation and handover execution. Just as FMIPv6 that supports all inter-domain handover scenarios, two modes (predictive and reactive) are defined in

> **Current BS**

**Predictive Mode:** Here, the current base station generates and broadcasts a Mobile Neighbor Advertisement (MOB\_NBR-ADV) message periodically. It contains the network topology and static link layer information. When the mobile node discovers a new base

FBack FBack

IEEE 802.16e network entry

Delivered Packets

FBU HI

UNA

HAck

Forward Packets

**PAR BS**

**NAR** 

DAD

**Target** 

preparation by sending an FBU message to the current access router.

mobile node sends an FNA message.

**NLD**

**MN L3** 

Fig. 3. FMIPv6 over IEEE 802.16e, Predictive Mode

**LSW** MOB-HO-IND

**LHI** MOB-BSHO-RSP

**MN L2** 

> MOB-MSHO-REQ PrRtAdv RtSolPr

Scanning

MOB-NBR-ADV

**LUP** 

RFC5270.

TN0G

TL2

TIND

TDEL

TL3

THI

station in this message, a scanning may be performed to acquire more dynamic parameters for the new base stations. If the newly found base stations are candidates for the target BSs, the NLD event is delivered to its IP layer from the mobile node MAC layer with the found BSIDs. The Router Solicitation Proxy (RtSolPr) message and Proxy Router Advertisement (PrRtAdv) messages are exchanged between the mobile node and previous access router. The terminal initiates handover by sending a Mobile Handover Request (MOB\_MSHO-REQ) message to the current base station and receives a Mobile Handover Response (MOB\_BSHO-RSP) message in reply with a target base station in it. The current base station may also initiate handover by sending a MOB\_BSHO-REQ message to the mobile node.

Fig. 4. FMIPv6 over IEEE 802.16e, Reactive Mode

After the mobile node receives MOB\_BSHO-REQ or MOB\_BSHO-RSP from the base station, the IP layer is triggered by link layer through a LHI to send Fast Binding Update (FBU) to the previous access router. The Handover Indication (HI) and Handover indication Acknowledge (Hack) messages are exchanged between previous and new access routers. The duplicate address detection is performed by new access router (it validates the uniqueness of NCoA in the new subnet, establishes tunnel and sends Fast Biding Acknowledge (FBack) message to the mobile station. Once the tunnel is established, the packets that are destined for the mobile node CoA are forwarded to the NCoA at the new access router through the tunnel. Upon receiving the FBack, the mobile node link layer is signalled by its network layer through a LSW to manage handover by sending a Mobile

Inter-Domain Handover in WiMAX Networks Using Optimized Fast Mobile IPv6 327

concept as the previous works. In the IFH802.16e, the previous access router is informed by

In this scheme, pre-established tunnelling mechanism to reduce handover preparation time is used. In addition, a set of messages has been defined to interleave layer 2 and layer 3 procedure. Cross layer design and cross function optimization are used to improve

In the OFHS, the serving base station periodically generates and sends the MOB\_NBR-ADV message to mobile stations. The MOB\_NBR-ADV message of IEEE 802.16e and the PrRtAdv message of FMIPv6 have similar functionality. The information of both messages can be sent through the MOB\_NBR-ADV message. Hence, these messages are merged and the PrSolPr message can be eliminated. The mobile station may also perform scanning to obtain link characteristics to evaluate whether to perform handover or otherwise. After the scanning procedure, mobile station selects target base stations among the candidate base stations,

base station to imitate IP layer handover on behalf of the mobile node.

Fig. 5. CLHS procedure

**3. Optimized Fast Handover Scheme (OFHS)** 

based on signal strength, QoS, service price and etc.

handover performance. The network model is as shown in Fig. 6.

handover indication (MOB-HO-IND) message to the target base station. This message starts the 802.16e network re-entry process. After re-entry process, the mobile node link layer triggers its network layer with a LUP to send Unsolicited Neighbor Advertisement (UNA) message to the new access router. When the new access router receives the UNA from the mobile node, it delivers the buffered packets to the mobile node.

**Reactive Mode:** If the mobile node sends the MOB-HO-IND message to the base station before receiving FBack, the mobile station carries out 802.16e network re-entry process without establishing tunnel with selected NAR. At this instant, the mobile node cannot perform predictive mode so it operates in reactive mode as follows. Upon the network entry procedure completion, the link layer of mobile node sends LUP signal to the IP layer. Then the IP layer identifies that it has moved to the target network without receiving the FBack in the previous link. The mobile node sends an UNA to the new access router by using NCoA as a source IP address and sends an FBU to the previous access router. When the new access router receives the UNA and the FBU from the mobile node, it sends the FBack to the previous access router, and the packets that have been forwarded from the previous access router to new access router are delivered to the mobile node (through NCoA) through the new access router.

#### **2.2.2 Cross Layer Handover Scheme (CLHS)**

(Chen & Hsieh, 2007) suggested an integrated design of layer 2 and layer 3 called Cross Layer Handover Scheme (CLHS). The main idea of the CLHS is that if the handover procedures of layer 2 and layer 3 can be coincident, the overall overhead of handover will be decreased. In the CLHS, the correlated messages of IEEE 802.16e and FMIPv6 were integrated. The authors show that some FMIPv6 handover information can be exchanged with the messages of IEEE 802.16e. The messages which have the same characteristics during handover procedure are merged. They are described as follows:

FBU-MOB\_HO-IND: The original MOB\_HO\_IND message are modified to include FBU as a new message. There are 6 reserve bits in the MOB\_HO\_IND message of link layer. One bit of them is used to indicate that the FBU is enabled or disabled. Upon receiving the FBU-MOB\_HO-IND message containing FBU bit, the current base station itself (instead of mobile node) sends FBU message to previous access router.

FNA\_RNG\_REQ: The RNG\_REQ message of 802.16e contains 8 reserved bits. They are used to send the information of FNA message of FMIPv6 in reactive mode.

In addition to the two messages, the neighbour advertisement message of layer 3 and the ranging request message of layer 2 were modified and merged. The MOB\_NBR\_ADV message in IEEE 802.16e and the PrRtAdv message in FMIPv6 have similar functionality. Hence, the CLHS merges these two messages together. The FBack massage of IP layer is combined with the Fast Ranging IE of link layer. Fig. 5 shows message sequence of the CLHS.

#### **2.2.3 Integrated fast handover in IEEE 802.16e (IFH802.16e)**

The IFH802.16e proposes a handover scheme for FMIPv6 over the IEEE 802.16e system by integrating FMIPv6 with IEEE 802.16e system. The IFH802.16e used same preparation concept as the previous works. In the IFH802.16e, the previous access router is informed by base station to imitate IP layer handover on behalf of the mobile node.

Fig. 5. CLHS procedure

326 Advanced Transmission Techniques in WiMAX

handover indication (MOB-HO-IND) message to the target base station. This message starts the 802.16e network re-entry process. After re-entry process, the mobile node link layer triggers its network layer with a LUP to send Unsolicited Neighbor Advertisement (UNA) message to the new access router. When the new access router receives the UNA from the

**Reactive Mode:** If the mobile node sends the MOB-HO-IND message to the base station before receiving FBack, the mobile station carries out 802.16e network re-entry process without establishing tunnel with selected NAR. At this instant, the mobile node cannot perform predictive mode so it operates in reactive mode as follows. Upon the network entry procedure completion, the link layer of mobile node sends LUP signal to the IP layer. Then the IP layer identifies that it has moved to the target network without receiving the FBack in the previous link. The mobile node sends an UNA to the new access router by using NCoA as a source IP address and sends an FBU to the previous access router. When the new access router receives the UNA and the FBU from the mobile node, it sends the FBack to the previous access router, and the packets that have been forwarded from the previous access router to new access router are delivered to the

(Chen & Hsieh, 2007) suggested an integrated design of layer 2 and layer 3 called Cross Layer Handover Scheme (CLHS). The main idea of the CLHS is that if the handover procedures of layer 2 and layer 3 can be coincident, the overall overhead of handover will be decreased. In the CLHS, the correlated messages of IEEE 802.16e and FMIPv6 were integrated. The authors show that some FMIPv6 handover information can be exchanged with the messages of IEEE 802.16e. The messages which have the same characteristics

FBU-MOB\_HO-IND: The original MOB\_HO\_IND message are modified to include FBU as a new message. There are 6 reserve bits in the MOB\_HO\_IND message of link layer. One bit of them is used to indicate that the FBU is enabled or disabled. Upon receiving the FBU-MOB\_HO-IND message containing FBU bit, the current base station itself (instead of mobile

FNA\_RNG\_REQ: The RNG\_REQ message of 802.16e contains 8 reserved bits. They are used

In addition to the two messages, the neighbour advertisement message of layer 3 and the ranging request message of layer 2 were modified and merged. The MOB\_NBR\_ADV message in IEEE 802.16e and the PrRtAdv message in FMIPv6 have similar functionality. Hence, the CLHS merges these two messages together. The FBack massage of IP layer is combined with the Fast Ranging IE of link layer. Fig. 5 shows message sequence of the

The IFH802.16e proposes a handover scheme for FMIPv6 over the IEEE 802.16e system by integrating FMIPv6 with IEEE 802.16e system. The IFH802.16e used same preparation

mobile node, it delivers the buffered packets to the mobile node.

mobile node (through NCoA) through the new access router.

during handover procedure are merged. They are described as follows:

to send the information of FNA message of FMIPv6 in reactive mode.

**2.2.3 Integrated fast handover in IEEE 802.16e (IFH802.16e)** 

**2.2.2 Cross Layer Handover Scheme (CLHS)** 

node) sends FBU message to previous access router.

CLHS.

#### **3. Optimized Fast Handover Scheme (OFHS)**

In this scheme, pre-established tunnelling mechanism to reduce handover preparation time is used. In addition, a set of messages has been defined to interleave layer 2 and layer 3 procedure. Cross layer design and cross function optimization are used to improve handover performance. The network model is as shown in Fig. 6.

In the OFHS, the serving base station periodically generates and sends the MOB\_NBR-ADV message to mobile stations. The MOB\_NBR-ADV message of IEEE 802.16e and the PrRtAdv message of FMIPv6 have similar functionality. The information of both messages can be sent through the MOB\_NBR-ADV message. Hence, these messages are merged and the PrSolPr message can be eliminated. The mobile station may also perform scanning to obtain link characteristics to evaluate whether to perform handover or otherwise. After the scanning procedure, mobile station selects target base stations among the candidate base stations, based on signal strength, QoS, service price and etc.

Inter-Domain Handover in WiMAX Networks Using Optimized Fast Mobile IPv6 329

station according to the policy and then carry out IEEE 802.16e network re-entry process. If the FBack is received by the mobile station before sending MOB-HO-IND message, handover continues in predictive mode. The MOB-HO-IND message contains selected target base station and the MAC address of the mobile station. The current base station notifies the new access router of the target base station by sending the HO-CONFIRM message. The previous access router obtains the exact target base station and related access router by receiving the HO-CONFIRM message. The previous access router starts forwarding the packets destined to the mobile station through one of the tunnels while the

The new access router buffers the packets during the network re-entry procedures. In this scheme layer 3 handover is initiated at the network side while the mobile station performs the layer 2 handover. Because the mobile station is not involved in formulating the NCoA, it should be informed of NCoA. This can be realized by sending the HO-COMPLETE message from target base station to the new access router after the network re-entry procedures of IEEE 802.16e. The target base station sends the REG-RSP message to mobile station and finalizes the network re-entry procedures of IEEE 802.16e and sends HO-COMPLETE message to confirm the layer 3 handover of mobile station. Upon HO-CONFIRM message received by the next access router, it starts delivering the buffered packets to the mobile station. The HO-COMPLETE message is necessary because after the mobile station performed layer 2 handover the NCoA should be notified to the mobile node. The new access router must send the Unsolicited Router Advertisement with Neighbor Advertisement Acknowledgement option to the mobile node. Fig. 7 shows OFHS predictive

other tunnels that are not selected are discarded.

Fig. 7. OFHS Handover Procedure, Predictive Mode

mode.

If handover is needed, the mobile station sends the MOB\_MSHO-REQ message to the possible target base stations that are listed. Then the current base station negotiates with the candidate base stations, and sends the recommended base stations and to mobile station through the MOB\_MSHO-RSP message. At the same time the current base station sends the handover notification (HO-NOTIF) message to previous access router. The HO-NOTIF message let the previous access router to start the layer 3 handover. It contains the identities of the recommended base stations and the MAC address of the mobile station. After receiving this message, the previous access router initiates the FMIPv6 handover by sending the handover initiate (HI) message to the next access router associated with target base station. The HI message should contain the NCoA of the mobile station when the stateless address auto-configuration (Thomson et al, 2007) is used. In the OFHS, the NCoA is configured by using the MAC address of the mobile node and the network prefix of new access router. It is performed by previous access router on behalf of the mobile station. The previous access router already knows the network prefix of new access router through some auxiliary protocols (Kwon et al., 2005; Liebsch et al., 2005).

Fig. 6. Network Model

The previous access router exchanges HI and handover acknowledge (HAck) messages with new access router. During this process, a tunnel between the previous and new access routers is set up and the validity of the NCoA is checked with duplicate address detection (DAD). The established tunnel may be more than one based on the recommended base stations. The tunnels are inactive and one of them will be activated only when previous access router receives the handover confirmation (HO-CONFRIM) message that includes the target base station. Once the tunnels are established, previous access router sends an FBack to the mobile station. FBack is applied to inform the status of the configuration of CoA. FBack are sent by the target base station so that the mobile station can be informed that the next CoA is valid. The mobile station can send a MOB-HO-IND message to the target base

If handover is needed, the mobile station sends the MOB\_MSHO-REQ message to the possible target base stations that are listed. Then the current base station negotiates with the candidate base stations, and sends the recommended base stations and to mobile station through the MOB\_MSHO-RSP message. At the same time the current base station sends the handover notification (HO-NOTIF) message to previous access router. The HO-NOTIF message let the previous access router to start the layer 3 handover. It contains the identities of the recommended base stations and the MAC address of the mobile station. After receiving this message, the previous access router initiates the FMIPv6 handover by sending the handover initiate (HI) message to the next access router associated with target base station. The HI message should contain the NCoA of the mobile station when the stateless address auto-configuration (Thomson et al, 2007) is used. In the OFHS, the NCoA is configured by using the MAC address of the mobile node and the network prefix of new access router. It is performed by previous access router on behalf of the mobile station. The previous access router already knows the network prefix of new access router through some

The previous access router exchanges HI and handover acknowledge (HAck) messages with new access router. During this process, a tunnel between the previous and new access routers is set up and the validity of the NCoA is checked with duplicate address detection (DAD). The established tunnel may be more than one based on the recommended base stations. The tunnels are inactive and one of them will be activated only when previous access router receives the handover confirmation (HO-CONFRIM) message that includes the target base station. Once the tunnels are established, previous access router sends an FBack to the mobile station. FBack is applied to inform the status of the configuration of CoA. FBack are sent by the target base station so that the mobile station can be informed that the next CoA is valid. The mobile station can send a MOB-HO-IND message to the target base

auxiliary protocols (Kwon et al., 2005; Liebsch et al., 2005).

Fig. 6. Network Model

station according to the policy and then carry out IEEE 802.16e network re-entry process. If the FBack is received by the mobile station before sending MOB-HO-IND message, handover continues in predictive mode. The MOB-HO-IND message contains selected target base station and the MAC address of the mobile station. The current base station notifies the new access router of the target base station by sending the HO-CONFIRM message. The previous access router obtains the exact target base station and related access router by receiving the HO-CONFIRM message. The previous access router starts forwarding the packets destined to the mobile station through one of the tunnels while the other tunnels that are not selected are discarded.

The new access router buffers the packets during the network re-entry procedures. In this scheme layer 3 handover is initiated at the network side while the mobile station performs the layer 2 handover. Because the mobile station is not involved in formulating the NCoA, it should be informed of NCoA. This can be realized by sending the HO-COMPLETE message from target base station to the new access router after the network re-entry procedures of IEEE 802.16e. The target base station sends the REG-RSP message to mobile station and finalizes the network re-entry procedures of IEEE 802.16e and sends HO-COMPLETE message to confirm the layer 3 handover of mobile station. Upon HO-CONFIRM message received by the next access router, it starts delivering the buffered packets to the mobile station. The HO-COMPLETE message is necessary because after the mobile station performed layer 2 handover the NCoA should be notified to the mobile node. The new access router must send the Unsolicited Router Advertisement with Neighbor Advertisement Acknowledgement option to the mobile node. Fig. 7 shows OFHS predictive mode.

Fig. 7. OFHS Handover Procedure, Predictive Mode

Inter-Domain Handover in WiMAX Networks Using Optimized Fast Mobile IPv6 331

In order to evaluate the performance of the proposed method, a numerical model has been developed. In this chapter, the important metrics for evaluating the handover mechanism are total handover procedure time and handover latency respectively. In the evaluation, the OFHS is compared with the RFC5270 as the reference procedure for using FMIPv6 in WiMAX.

To analyze the performance model of the proposed scheme, the duration of each part of the handover procedure are considered. The message interaction is based on the duration of a frame which is an OFDMA type used by IEEE 802.16e air interface. The frame duration is assumed to be at least 1ms and processing time is ignored since it is less than the frame duration. On the other hand, the network nodes message transmission delay is at least a frame long (>1ms). The radio propagation delay is assumed to be smaller than the frame

The total handover procedure time (*TTHT*) is defined as the elapsed time between a mobile node sending the MOB\_MSHO-REQ message to the current base station and the time the mobile station can receive the first packet through the target access router. *TTH-PM-RFC* and *TTH-PM-OFHS* are defined as the total handover time of the predictive mode in RFC5270 and OFHS, respectively. The Equations are defined in term of delay of every routing hop in a wired backbone (*THOP*) and frame duration of IEEE 802.16e (*TF*). Negotiation between the current base station and the target base station is started by sending MOB-MSHO-REQ. Then the current base station sends handover notification message to target base station and receives handover notification response from it. The procedure is concluded by sending MOB-BS-HO-RSP to current base station. The time lag from the point of sending MOB-MSHO-REQ to receiving MOB-BSHO-RSP or negotiation delay between the current and recommended base stations (*TNEG*) is given by Equation (1). The time required to perform FMIPv6 in layer 3 from the point of sending FBU to receiving FBack is *TL3* and the latency of IEEE 802.16e network re-entry procedure is given by *TL2*.They are expressed in Equations (2) and (3), respectively. *NPAR-NAR* is the distance between the previous and new access routers in term of number of hops and *TDAD* is time needed to complete a duplicate address detection procedure. The MAC layer handover time is based on the number of messages exchanged between mobile station and base stations according to the RFC5270. Packet delivery time (*TDEL*) is the time required from the point of sending the UNA message after IEEE 802.16e handover to receiving the first packet from new access router; this is given by

TNEG=4THOP+2NPAR-NAR×THOP (1)

TL3-RFC=3TF + 2THOP+2NPAR-NAR × THOP + TDAD (2)

TL2= 10TF + 30 (ms) (3)

 TDEL-RFC= 3TF + 2THOP (4) The elapse time between receiving MOB-BSHO-RSP and starting layer 3 handover is given by *THI* (For RFC5270 procedure *THI = 2TF*). *TIND* is elapse time between receiving FBack and

sending MOB-HO-IND. To simplify analysis, fixed delay time for *TIND* is assumed.

**4. Performance evaluation** 

duration, so it is omitted.

Equation (4).

**4.1 Total handover procedure time** 

If the mobile node sends MOB-HO-IND message to the current base station before receiving FBack (before establishing tunnel with selected access router), the mobile station starts IEEE 802.16e network re-entry process and the current base station sends HO-CONFIRM message to the previous access router. The previous access router stops sending packets to the mobile node and starts to buffer the packets destined for the mobile station. During the network reentry procedures of IEEE 802.16e or after that, the previous access router receives the HAck message. There are two scenarios; first, if the previous access router receives HAck messages from the new access router before the end of network re-entry procedures of IEEE 802.16e, the previous access router starts to tunnel the packets destined for the current CoA to the new CoA at the new access router. Then the new access router starts delivering the packets to the mobile station. The previous access router already knows the exact target mobile station and its associated access router, therefore, the previous access router can determine through which tunnel it should start forwarding the packets destined to the mobile station while the other tunnels that are not used will be discarded. The second scenario is that, if the network re-entry procedure of IEEE 802.16e is terminated and the tunnel with selected new access router has not been established yet, the previous access router waits to receive HAck message from the new access router. Upon receiving the HAck message, the previous access router starts to tunnel the packets (destined for the current CoA) to the NCoA at the new access router. Then the new access router starts delivering the tunnelled packets to the mobile node. These two scenarios are called semi-predictive mode defined in OFHS instead of reactive mode defined in the RFC5270. The semi-predictive mode procedure is shown in Fig. 8.

Fig. 8. OFHS Handover Procedure, Semi-Predictive Mode

#### **4. Performance evaluation**

330 Advanced Transmission Techniques in WiMAX

If the mobile node sends MOB-HO-IND message to the current base station before receiving FBack (before establishing tunnel with selected access router), the mobile station starts IEEE 802.16e network re-entry process and the current base station sends HO-CONFIRM message to the previous access router. The previous access router stops sending packets to the mobile node and starts to buffer the packets destined for the mobile station. During the network reentry procedures of IEEE 802.16e or after that, the previous access router receives the HAck message. There are two scenarios; first, if the previous access router receives HAck messages from the new access router before the end of network re-entry procedures of IEEE 802.16e, the previous access router starts to tunnel the packets destined for the current CoA to the new CoA at the new access router. Then the new access router starts delivering the packets to the mobile station. The previous access router already knows the exact target mobile station and its associated access router, therefore, the previous access router can determine through which tunnel it should start forwarding the packets destined to the mobile station while the other tunnels that are not used will be discarded. The second scenario is that, if the network re-entry procedure of IEEE 802.16e is terminated and the tunnel with selected new access router has not been established yet, the previous access router waits to receive HAck message from the new access router. Upon receiving the HAck message, the previous access router starts to tunnel the packets (destined for the current CoA) to the NCoA at the new access router. Then the new access router starts delivering the tunnelled packets to the mobile node. These two scenarios are called semi-predictive mode defined in OFHS instead of reactive mode defined in the RFC5270. The semi-predictive mode procedure is shown in

Fig. 8.

Fig. 8. OFHS Handover Procedure, Semi-Predictive Mode

In order to evaluate the performance of the proposed method, a numerical model has been developed. In this chapter, the important metrics for evaluating the handover mechanism are total handover procedure time and handover latency respectively. In the evaluation, the OFHS is compared with the RFC5270 as the reference procedure for using FMIPv6 in WiMAX.

To analyze the performance model of the proposed scheme, the duration of each part of the handover procedure are considered. The message interaction is based on the duration of a frame which is an OFDMA type used by IEEE 802.16e air interface. The frame duration is assumed to be at least 1ms and processing time is ignored since it is less than the frame duration. On the other hand, the network nodes message transmission delay is at least a frame long (>1ms). The radio propagation delay is assumed to be smaller than the frame duration, so it is omitted.

#### **4.1 Total handover procedure time**

The total handover procedure time (*TTHT*) is defined as the elapsed time between a mobile node sending the MOB\_MSHO-REQ message to the current base station and the time the mobile station can receive the first packet through the target access router. *TTH-PM-RFC* and *TTH-PM-OFHS* are defined as the total handover time of the predictive mode in RFC5270 and OFHS, respectively. The Equations are defined in term of delay of every routing hop in a wired backbone (*THOP*) and frame duration of IEEE 802.16e (*TF*). Negotiation between the current base station and the target base station is started by sending MOB-MSHO-REQ. Then the current base station sends handover notification message to target base station and receives handover notification response from it. The procedure is concluded by sending MOB-BS-HO-RSP to current base station. The time lag from the point of sending MOB-MSHO-REQ to receiving MOB-BSHO-RSP or negotiation delay between the current and recommended base stations (*TNEG*) is given by Equation (1). The time required to perform FMIPv6 in layer 3 from the point of sending FBU to receiving FBack is *TL3* and the latency of IEEE 802.16e network re-entry procedure is given by *TL2*.They are expressed in Equations (2) and (3), respectively. *NPAR-NAR* is the distance between the previous and new access routers in term of number of hops and *TDAD* is time needed to complete a duplicate address detection procedure. The MAC layer handover time is based on the number of messages exchanged between mobile station and base stations according to the RFC5270. Packet delivery time (*TDEL*) is the time required from the point of sending the UNA message after IEEE 802.16e handover to receiving the first packet from new access router; this is given by Equation (4).

$$\mathbf{T}\_{\rm NEG} = \mathbf{4}\mathbf{T}\_{\rm HOP} + \mathbf{2}\mathbf{N}\_{\rm PAB}\mathbf{\_{NAR}} \times \mathbf{T}\_{\rm HOP} \tag{1}$$

$$\mathbf{T\_{L3:RFC}} = \mathbf{3}\mathbf{T\_F} + \mathbf{2}\mathbf{T\_{HCP}} + \mathbf{2}\mathbf{N\_{PAR:NAR}} \times \mathbf{T\_{HCP}} + \mathbf{T\_{DAD}}\tag{2}$$

$$\mathrm{T\_{12} = 10T\_F + \text{\textquotedblleft O3 (ms)}}\tag{3}$$

$$\mathbf{T}\_{\text{DEL-HFC}} = \mathbf{\mathcal{ST}}\_{\text{F}} + \mathbf{\mathcal{T}}\_{\text{HCP}} \tag{4}$$

The elapse time between receiving MOB-BSHO-RSP and starting layer 3 handover is given by *THI* (For RFC5270 procedure *THI = 2TF*). *TIND* is elapse time between receiving FBack and sending MOB-HO-IND. To simplify analysis, fixed delay time for *TIND* is assumed.

Inter-Domain Handover in WiMAX Networks Using Optimized Fast Mobile IPv6 333

router. After the previous access router sends the FBAck message to the mobile node, it stops delivering packets to the CoA (sending packets to mobile node). At this time, the current access router re-routes the packets that destined to the CoA to the NCoA in the target access router. Hence, the actual period of handover latency in predictive mode begins when the mobile node receives an FBAck message. In reactive mode, the actual period of the handover latency begins by sending the MOB-HO-IND message. *THL-PM-RFC* is defined as the handover latency of the predictive mode of the RFC 5270 and *THL-PM-OFHS* as the handover

 THL-PM-OFHS =TL2+TDEL=11TF+30(ms) + [THOP]F(11) *THL-RM-RFC* is the handover latency of the reactive mode of the RFC5270 and *THL-SPM-OFHS* is the total handover latency of the semi-predictive mode of the OFHS given by Equations (12)

THL-RM-RFC= TL2 + TFNA+TL3-RM +TDEL (12)

THL-SPM-OFHS = TL2+ TDEL + T'IND (13)

The parameters of OFHS and RFC570 are compared in this section, based on the previous

Parameter Value

*THOP* 1 ms

*N PAR-NAR* 2 hops

*TDAD* 800 ms

*THI = TIND TF*

Fig. 9 shows total handover time of the RFC5270 and OFHS in term of frame durations for predictive, semi-predictive and reactive modes according to Equations (6) to (9), respectively. Handover latency variation in term of frame duration for all modes of the

+[*NPAR-NAR×THOP*]*F +* [*THOP*]*F* 

THL-PM-RFC = TIND + TL2 + TDEL-RFC (10)

= *11TF+30(ms)+*[*2THOP +2 NPAR-NAR×THOP*]*<sup>F</sup>*

= *11TF+30(ms)+*[*THOP*]*<sup>F</sup>* + *T'IND* 

= TIND +14TF +30(ms) + [2THOP]F

latency of predictive mode of OFHS given by Equations (10) and (11), respectively.

and (13), respectively.

**5. Results and discussion** 

Table 1. Network Parameters

analysis. Handover parameters are given as in Table 1.

The message interaction is based on the duration of a frame, all times expressed as integer number of frame. Therefore, all non-integer times is rounded to the next nearest integer number (this is shown as [ ]*F*). In OFHS, *TNEG*, *TL2* and *TDEL* are the same as Equations (1), (2) and (3), and *TL3* is obtained from Equations (5). Hence, the total handover time of the predictive mode in term of *TF* for RFC5270 and OFHS are given by Equation (6) and (7), respectively.

$$\rm T\_{\rm L3,OHFS} = \rm T\_{\rm F} + 2\ \rm T\_{\rm HOP} + 2\ \rm N\_{\rm PAR,NAR} \times \rm T\_{\rm HOP} + \rm T\_{\rm DAD} \tag{5}$$

$$\rm T\_{\rm H+\rm PMR,BF} = \rm T\_{\rm NEG} + \rm T\_{\rm H+} + \rm T\_{\rm L3,BF} + \rm T\_{\rm D2+} + \rm T\_{\rm D2+} \rm R} \tag{6}$$

$$= [\rm 4T\_{\rm HOP} + 2\ \rm N\_{\rm PAR,NR} \times \rm T\_{\rm HOP}] + 18\ \rm T\_{\rm F} + [\rm 2T\_{\rm HOP}] + \rm \tag{7}$$

$$+ [2\ \rm T\_{\rm HOP} + 2\ \rm N\_{\rm PAR,NR} \times \rm T\_{\rm HOP} + \rm T\_{\rm DAD}] \tag{7}$$

$$\rm T\_{\rm H+\rm PMR,OFS} = \rm T\_{\rm NEG} + \rm T\_{\rm L3,FOR} + \rm T\_{\rm HND} + \rm T\_{\rm D2+} \rm T\_{\rm D2+} \rm R} \tag{7}$$

$$= [\rm 4T\_{\rm HOP} + 2\ \rm N\_{\rm PAR,NR} \times \rm T\_{\rm HOP}] + 12\ \rm T\_{\rm F} +$$

$$+ [2\ \rm T\_{\rm HOP} + 2\ \rm N\_{\rm PAR,NR} \times \rm T\_{\rm HOP} + \rm T\_{\rm DAD}] \mighth$$

*TTH-RM-RFC* is the total handover time of the reactive mode of the RFC 5270 and *TTH-SPM-OFHS* as the total handover time of the semi-predictive mode of the OFHS given by Equations (8) and (9) respectively. In reactive mode, after sending FBU, the mobile node does not receive an FBAck from the current access router before the mobile node is forced to move to the target access router. The mobile station must wait for packet rerouting before it can receive any packets from the target access router. *TFNA* is elapse time between layer 2 handover termination and FNA message, and the time required performing FMIPv6 L3 handover from sending *FBU* to mobile node receiving *FBack* is *TL3-RM*. In reactive mode and semipredictive mode *TIND* has various values depending on location, direction and speed of mobile station. Also, *TDEL* depends on the number of buffered packets and frame duration.

$$\begin{aligned} \mathbf{T\_{TH-RM-RFC}} &= \mathbf{T\_{NHG}} + \mathbf{T\_{HI}} + \mathbf{T\_{IND}} + \mathbf{T\_{12}} + \mathbf{T\_{FAA}} + \mathbf{T\_{13-RM}} + \mathbf{T\_{DEL-RFC}} \\ &= \mathbf{T\_{NEG}} + \mathbf{T\_{HI}} + \mathbf{T\_{ND}} + \mathbf{T\_{12}} + \mathbf{T\_{FAA}} + \mathbf{T\_{13-RM}} + \mathbf{T\_{DEL-RFC}} \\ &= [2\mathbf{T\_{HOP}} + 2\mathbf{N\_{PAR-NR}} \times \mathbf{T\_{HOP}}]\mathbf{F} + \mathbf{T\_{NDI}} + 13\mathbf{T\_F} + 30\mathbf{(ms)} + \\ &+ [\mathbf{T\_{HOP}}]\_F + [2\mathbf{N\_{PAR-NR}} \times \mathbf{T\_{HOP}}]\_F + [\mathbf{N\_{PAR-MR}} \times \mathbf{T\_{HOP}}]\_F + [\mathbf{T\_{HOP}}]\_F \\ &\quad \mathbf{T\_{THS-RM}} \mathbf{G\_{TH}} = \mathbf{T\_{NEG}} + \mathbf{T\_{NDI}} + \mathbf{T\_{12}} + \mathbf{T\_{DEL-PRO}} \\ &= [2\mathbf{T\_{HOP}} + 2\mathbf{N\_{PAR-NR}} \times \mathbf{T\_{HOP}}]\_F + \mathbf{T\_{NDI}} + \\ &\quad + 11\mathbf{T\_F} + 30\mathbf{(ms)} + 2 \ [\mathbf{T\_{HOP}}]\_F \end{aligned} \tag{9}$$

#### **4.2 Handover latency**

Handover latency (*THL*) is defined as the elapsed time between a mobile node receiving the last packet through its current access router and the first packet through the target access router. After the previous access router sends the FBAck message to the mobile node, it stops delivering packets to the CoA (sending packets to mobile node). At this time, the current access router re-routes the packets that destined to the CoA to the NCoA in the target access router. Hence, the actual period of handover latency in predictive mode begins when the mobile node receives an FBAck message. In reactive mode, the actual period of the handover latency begins by sending the MOB-HO-IND message. *THL-PM-RFC* is defined as the handover latency of the predictive mode of the RFC 5270 and *THL-PM-OFHS* as the handover latency of predictive mode of OFHS given by Equations (10) and (11), respectively.

$$\mathcal{T}\_{\text{HL-PM-RFC}} = \mathcal{T}\_{\text{IND}} + \mathcal{T}\_{\text{L2}} + \mathcal{T}\_{\text{DEL-RFC}} \tag{10}$$

$$= \mathcal{T}\_{\text{IND}} + \text{14}\mathcal{T}\_{\text{F}} + \text{30(ms)} + \text{[2T}\_{\text{HOF}}]\_{\text{F}}$$

$$\mathbf{T\_{HL-FM-OFFS}} = \mathbf{T\_{L2}} + \mathbf{T\_{DHL}} = \mathbf{11T\_F} + \mathbf{30(ms)} + \mathbf{[T\_{HCP}]\_F} \tag{11}$$

*THL-RM-RFC* is the handover latency of the reactive mode of the RFC5270 and *THL-SPM-OFHS* is the total handover latency of the semi-predictive mode of the OFHS given by Equations (12) and (13), respectively.

$$\rm T\_{HL-RM.HG} = T\_{12} + T\_{FNA} + T\_{13.RM} + T\_{IDL} \tag{12}$$

$$= 117 \pm 30 \text{(ms)} + [2T\_{HOP} + 2 \text{ } N\_{PAR-MAR} \times T\_{HOP}]\_F$$

$$+ [\rm N\_{PAR.MAR} \times T\_{HOP}]\_F + [\rm T\_{HOP}]\_F$$

$$\rm T\_{HL-SFR} = T\_{12} + \rm T\_{DL} + \rm T'\_{IND} \tag{13}$$

$$= 117 \pm 30 \text{(ms)} + [T\_{HOP}]\_F + \rm T'\_{ND}$$

#### **5. Results and discussion**

332 Advanced Transmission Techniques in WiMAX

The message interaction is based on the duration of a frame, all times expressed as integer number of frame. Therefore, all non-integer times is rounded to the next nearest integer number (this is shown as [ ]*F*). In OFHS, *TNEG*, *TL2* and *TDEL* are the same as Equations (1), (2) and (3), and *TL3* is obtained from Equations (5). Hence, the total handover time of the predictive mode in term of *TF* for RFC5270 and OFHS are given by Equation (6) and (7),

TL3-OFHS = TF +2 THOP+2 NPAR-NAR × THOP +TDAD (5)

TTH-PM-RFC =TNEG +THI +TL3-RFC +TIND+TL2 +TDEL-RFC (6)

+TIND +30(ms) + [THOP]F *TTH-RM-RFC* is the total handover time of the reactive mode of the RFC 5270 and *TTH-SPM-OFHS* as the total handover time of the semi-predictive mode of the OFHS given by Equations (8) and (9) respectively. In reactive mode, after sending FBU, the mobile node does not receive an FBAck from the current access router before the mobile node is forced to move to the target access router. The mobile station must wait for packet rerouting before it can receive any packets from the target access router. *TFNA* is elapse time between layer 2 handover termination and FNA message, and the time required performing FMIPv6 L3 handover from sending *FBU* to mobile node receiving *FBack* is *TL3-RM*. In reactive mode and semipredictive mode *TIND* has various values depending on location, direction and speed of mobile station. Also, *TDEL* depends on the number of buffered packets and frame duration.

TTH-RM-RFC =TNEG +THI +TIND+TL2+TFNA +TL3-RM +TDEL-RFC (8)

TTH-SPM-OFHS = TNEG +TIND+ TL2 + TDEL-PRO (9)

Handover latency (*THL*) is defined as the elapsed time between a mobile node receiving the last packet through its current access router and the first packet through the target access

*=TNEG +THI +TIND+TL2 +TFNA +TL3-RM +TDEL-RFC* 

*+11TF +30(ms) + 2* [*THOP*]*F* 

*=* [*2THOP +2NPAR-NAR ×THOP*]*F +TIND+ 13TF +30(ms)+* 

+[*THOP*]*F +*[*2NPAR-NAR ×THOP*]*F +*[*NPAR-NAR ×THOP*]*F +* [*THOP*]*F* 

*=* [*2THOP +2NPAR-NAR×THOP*]*F +TIND+*

=[4THOP +2 NPAR-NAR ×THOP]F +18TF + [2THOP]F +

+[2THOP +2NPAR-NAR ×THOP +TDAD]F +TIND +30(ms)

TTH-PM-OFHS =TNEG +TL3-POR +TIND+TL2 +TDEL-POR (7)

= [4THOP +2NPAR-NAR ×THOP]F+12TF+

+[2THOP +2NPAR-NAR× THOP +TDAD]F+

respectively.

**4.2 Handover latency** 

The parameters of OFHS and RFC570 are compared in this section, based on the previous analysis. Handover parameters are given as in Table 1.


Table 1. Network Parameters

Fig. 9 shows total handover time of the RFC5270 and OFHS in term of frame durations for predictive, semi-predictive and reactive modes according to Equations (6) to (9), respectively. Handover latency variation in term of frame duration for all modes of the

Inter-Domain Handover in WiMAX Networks Using Optimized Fast Mobile IPv6 335

semi-predictive and reactive mode compare with RFC5270. The reason is that our scheme needs less number of messages than that of the RFC5270 when performing handover, and pre-established tunnel concept prepare a mechanism to reduce handover time. Also, the additional anticipation time imposed by FMIPv6 that causes the handover execution start earlier than planned is solved. In OFHS, occurrence probability of reactive mode is lower than that of the RFC5270, because earlier handover preparation provides sufficient time for

In this chapter an overview of inter-domain handover in WiMAX networks have been presented. The previous solutions for applying FMIPv6 on IEEE 802.16e have long latency that are not acceptable for real time services such as video streaming and voice over IP. In order to reduce handover latency, an optimized fast IPv6 handover scheme (OFHS) have been proposed. The OFHS combined cross layer design and cross function optimization to achieve lower handover latency. A pre-established multi tunnelling concept and a buffered routers mechanism have used to prepare seamless handover. The Layer 2 handover in 802.16e and layer 3 handover in FMIPv6 procedures are interleaved and the correlated

The results show that OFHS reduces handover latency and packet losses, and increase probability of predictive mode that has lower handover latency than reactive mode compared with RFC5270. The OFHS reduces handover latency by 38.2% in predictive mode.

Chen, Y. and Hsieh, F. (2007). A Cross Layer Design for Handoff in 802.16e Network with IPv6 Mobility, *IEEE Communications Society subject matter experts*, 2007 Han, Y.; Jang, H.; Choi, J.; Park, B. and McNair, J. (2007). A Cross-Layering Design forIPv6

IEEE 802.16e (2004). IEEE Standard for Local and Metropolitan Area Networks, part 16: Air Interface for Fixed and Mobile Broadband Wireless Access Systems*, IEEE* 

Fast Handover Support in an IEEE 802.16e Wireless MAN*, IEEE Network*, Dec 2007.

the mobile node to receive FBack and drive predictive mode.

Fig. 11. Handover Latency for Ordinary Frame Duration (5ms)

messages for both layers are blended and reconstructed, effectively.

**6. Chapter summary** 

**7. References** 

RFC5270 and the OFHS are depicted in Fig. 10. The numerical values are obtained from Equations (10) to (13). Fig. 8 and Fig. 9 show that, the delay increases with the frame duration increases. The reason is that the base station replies the received message at the next frame because the current frame resource utilization is scheduled in advance. Additionally the response time is lengthened as the frame duration increases. The OFHS shows better total handover time and handover latency than RFC5270.

Fig. 9. Total Handover time versus Frame Duration

Fig. 10. Handover Latency versus Frame Duration

Usually in IEEE 802.16e, frame duration is considered as 5ms. In Fig. 11 total handover time and handover latency of RFC5270 and OFHS in reactive, predictive and semi-predictive modes for 5ms frame duration are illustrated. When frame duration is 5ms, OFHS decreases total handover time to 47ms for predictive mode, and 90ms for semi-predictive and reactive mode. The OFHS also reduces handover latency to 47ms for predictive mode, and 672ms for semi-predictive and reactive mode compare with RFC5270. The reason is that our scheme needs less number of messages than that of the RFC5270 when performing handover, and pre-established tunnel concept prepare a mechanism to reduce handover time. Also, the additional anticipation time imposed by FMIPv6 that causes the handover execution start earlier than planned is solved. In OFHS, occurrence probability of reactive mode is lower than that of the RFC5270, because earlier handover preparation provides sufficient time for the mobile node to receive FBack and drive predictive mode.

Fig. 11. Handover Latency for Ordinary Frame Duration (5ms)

#### **6. Chapter summary**

334 Advanced Transmission Techniques in WiMAX

RFC5270 and the OFHS are depicted in Fig. 10. The numerical values are obtained from Equations (10) to (13). Fig. 8 and Fig. 9 show that, the delay increases with the frame duration increases. The reason is that the base station replies the received message at the next frame because the current frame resource utilization is scheduled in advance. Additionally the response time is lengthened as the frame duration increases. The OFHS

shows better total handover time and handover latency than RFC5270.

Fig. 9. Total Handover time versus Frame Duration

Fig. 10. Handover Latency versus Frame Duration

Usually in IEEE 802.16e, frame duration is considered as 5ms. In Fig. 11 total handover time and handover latency of RFC5270 and OFHS in reactive, predictive and semi-predictive modes for 5ms frame duration are illustrated. When frame duration is 5ms, OFHS decreases total handover time to 47ms for predictive mode, and 90ms for semi-predictive and reactive mode. The OFHS also reduces handover latency to 47ms for predictive mode, and 672ms for In this chapter an overview of inter-domain handover in WiMAX networks have been presented. The previous solutions for applying FMIPv6 on IEEE 802.16e have long latency that are not acceptable for real time services such as video streaming and voice over IP. In order to reduce handover latency, an optimized fast IPv6 handover scheme (OFHS) have been proposed. The OFHS combined cross layer design and cross function optimization to achieve lower handover latency. A pre-established multi tunnelling concept and a buffered routers mechanism have used to prepare seamless handover. The Layer 2 handover in 802.16e and layer 3 handover in FMIPv6 procedures are interleaved and the correlated messages for both layers are blended and reconstructed, effectively.

The results show that OFHS reduces handover latency and packet losses, and increase probability of predictive mode that has lower handover latency than reactive mode compared with RFC5270. The OFHS reduces handover latency by 38.2% in predictive mode.

#### **7. References**


IEEE 802.21 (2009). IEEE Standard for Local and metropolitan area networks- Part 21: Media

Jang, H. J.; Jee, J. ; Han, Y.H. ; Park, S.D. and Cha., J. (2008). Mobile IPv6 Fast Handovers

Johnson, D.; Perkins. C. and Arkko J. (2004). Mobility Support in IPv6, *RFC 3775, Internet* 

Koodli, R. (2005). Fast Handovers for Mobile IPv6, *RFC 4068, Internet Engineering Task Force,* 

Kwon, D. H.; Kim, Y. S.; Bae, K. J. and Suh, Y. J. (2005). Access router information protocol

Lee, D.H.; Kyamakya, K. and Umondi, J.P. (2006). *Fast handover algorithm for IEEE 802.16e* 

Lee, J.S.; Choi, S.Y. and Eom, Y.I. (2009). Fast Handover Scheme Using Temporary CoA in

Liebsch, M.; Ed.; Singh, A.; Chaskar, H.; Funato, D. and Shim, E. (2005). Candidate Access Router Discovery (CARD), *RFC 4066, Internet Engineering Task Force*, 2005. Park, J.; Kwon, D.; Suh, Y. (2006). An Integrated Handover Scheme for Fast Mobile IPv6 over

Perkins, C. (2002). IP Mobility Support for IPv4, *RFC 3344, Network Working Group of Internet* 

Seyyedoshohadaei, S.M.; Khatun S.; Mohd Ali, B.; Othman, M. and Anwar, F. (2009). An

Seyyedoshohadaei, S.M.; Mohd Ali, B.; Othman, M. and Khatun S. (2011). Network Mobility

Thomson, S.; Narten, T. ; and Jinmei, T. (2007). IPv6 Stateless Address Auto configuration, *RFC 4862, Network Working Group of Internet Engineering Task Force,* Sep 2007

Integrated Scheme to Improve Performance of Fast Mobile IPv6 Handover in IEEE 802.16e Network, *proceeding of MICC 2009 Malaysian International Conference on* 

in IEEE 802.16e Network using Fast Mobile IPv6, *proceeding of ASME International Conference on Communication and Broadband Networking ICCBN 2011,* pp.747-752,

over IEEE 802.16e Networks, *RFC5270 of Internet Engineering Task Force, Network* 

with FMIPv6 for efficient handovers and their implementations, *Globecom*, pp.

*Broadband Wireless Access System*, *In Wireless Pervasive Computing*, *1st International* 

Mobile WiMAX Systems, *In 11th International Conference in Advanced Communication* 

Independent Handover, *IEEE*

*Network Working Group,* July2005

*Technology, ICACT 2009*, pp. 1772–1776 (2009)

*Communication,* Kuala Lumpur, Malaysia, Des 2009

*Working Group*, Mar. 2008.

*Engineering Task Force* 

3814-3819, 2005

*symposium,* Jan. 2006

IEEE 802.16e Systems, 2006

*Engineering Task Force,* July 2005.

Kuala Lumpur, Malaysia, Jun 2011
