**1. Introduction**

34 Will-be-set-by-IN-TECH

[70] Fossorier M. P. C., Mihaljevic M., et al. (1999) Reduced complexity iterative decoding of low-density parity check codes based on belief propagation. Communications, IEEE

Transactions on, vol. 47, pp. 673-680, 1999.

The use of ultra wide band (UWB) signals can offer many advantages for communications. It can provide a very robust performance even under harsh multipath and interference conditions, the capability of precision ranging and a reduced power consumption. Since the power spectral density is very low, it is possible to overlay UWB networks with already existing non-UWB emissions.

Early UWB concepts for communications have almost exclusively relied on impulse radio, where the whole available bandwidth, i.e., up to 7.5 GHz, is covered at once by means of very short pulses which are generated with a low duty cycle. Meanwhile, a bandwidth of 7.5 GHz is only available in the US [1, 3]. In Europe, the spectrum which is available with the same transmit power spectral density of -41.3 dBm/MHz ranges only from 6.0 to 8.5 GHz [4], if no detect and avoid techniques are applied1. A potential UWB system has therefore to be able to 'live' with a mean transmit power of less than -7.3 dBm.

This is a small value, but fortunately UWB systems may exploit the signal energy very efficiently because firstly, even at data rates in the Gbps range, it is not required to use bandwidth efficient (but energy inefficient) modulation schemes like a 1024-QAM. Secondly, UWB transmission benefits from a good fading resistance.

For a measured indoor channel [7], Fig. 1 shows that even a bandwidth of 'only' 500 MHz ensures a very good fading resistance: If the receiver is moved over all *x*-*y*-positions in the non-LOS case, the smallest power value at the receive antenna lies less than 3 dB below the mean power, averaged over all positions. Thus the fading margin could be chosen in the order of 3-4 dB — even for indoor channels which exhibit the largest coherence bandwidth.

The second energy efficiency argument claimed above is underpinned in Fig. 2. It shows the channel capacity depending on the bandwidth, where additive white Gaussian noise (AWGN) is assumed. A value of 83 MHz just corresponds to the total bandwidth available in the 2.4 GHz ISM band, which is chosen for comparison. At 1 Gbps, a 2.5 GHz bandwidth promises an advantage of 25 dB with respect to the required receive-power. Furthermore, even binary modulation (on the inphase and quadrature components) promises high data rates.

<sup>1</sup> With detect and avoid techniques, -41.3 dBm/MHz is also permitted between 4.2 and 4.8 GHz.

©2013 Song et al., licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0),which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. ©2013 Song et al., licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

#### 2 Will-be-set-by-IN-TECH 110 Ultra-Wideband Radio Technologies for Communications, Localization and Sensor Applications Non-Coherent UWB Communications <sup>3</sup>

Unfortunately, a very large signal bandwidth is also associated with some serious problems. These problems are related to the transceiver components itself (availability of broadband antennas, amplifiers etc.), and to the technical effort which is required for synchronization, channel estimation and interference rejection. Since UWB networks operate in frequency bands already assigned to other RF-systems, the probability that narrowband interference occurs at all increases with the bandwidth, too.

Furthermore, by increasing the bandwidth, more and more multipath arrivals with different path gains and delays are resolvable at the receiver, which makes it more difficult to collect the multipath energy coherently — although the power at the receive antenna does not suffer from the fading effect. Fig. 1 shows an example that one may lose 10 dB and more, if an UWB-receiver uses only the strongest signal echo. Thus, especially in non-LOS scenarios, a coherent RAKE receiver requires a very large number of RAKE fingers and a precise channel knowledge to efficiently capture the multipath energy. Such a coherent RAKE receiver will be very complex and costly, such that the hardware itself may consume a lot of power. This fact is the major motivation for systems using non-coherent detection, which are discussed in this chapter.

−100 −95 −90 −85 −80 −75 −70 −65 −60

non-coherent receiver is clearly the dramatically reduced effort which is required for channel estimation, synchronization, and multipath diversity combining. This advantage is, however, bought by a serious drawback: non-coherent detectors are more susceptible to narrowband

Non-coherent detection can either rely on envelope detection or on differential detection. In the simplest case, path-diversity combining is carried out via an analog integration device. However, the change from analog to digital combining stimulates new perspectives. Since a digital code matched filter can be applied prior to the non-coherent part of the receiver, the capability to distinguish users (or networks) by means of code division multiple access is improved. We show that digital receiver implementations with user specific filtering have also an enhanced interference rejection capability and energy efficiency. Moreover, we present well suited solutions for the analog-to-digital conversion, the spread-spectrum code sequences,

Although non-coherent detection is not restricted to low data rates — even orthogonal frequency division multiplex (OFDM) can be combined with non-coherent modulation and

Non-coherent detection can either be based on envelope detection or on differential detection. In the simplest case, path-diversity combining can be achieved by means of a single, analog integrate and dump filter, see Fig. 3 and Fig. 4. The integration effectively provides a binary weighting of the multipath arrivals: all components inside the integration window of size *T*int are weighted with "1", while all the others are weighted with "0". Regardless of whether envelope or differential detection is chosen, we assume that the receiver uses a quadrature

detection [19] — we focus our attention on low data rate single carrier transmission.

**Figure 2.** Capacity of an AWGN channel as a function of the receive-power for different signal

interference (NBI), multi-user interference (MUI), and inter-symbol interference (ISI).

bandwidths. The 2.4 GHz ISM band offers a bandwidth of 83 MHz.

**2. Non-coherent detection in multipath AWGN**

Rx−power in dBm

B=83 MHz B=500 MHz B=1000 MHz B=2500 MHz binary modulation

Non-Coherent UWB Communications 111

SISO AWGN channel, Nyquist rate

101

and the modulation format.

102

channel capacity in Mbps

103

104

**Figure 1.** Received energy *E*rx normalized by the transmitted symbol's energy *E*tx in dB, versus different signal bandwidths. The thickness of the curves indicates LOS or non-LOS regimes. The curves without markers show *E*rx/*E*tx averaged across all *x*-*y*-positions within an rectangular area of 30 cm × 40 cm (1 cm grid, data from [7]), if an ideal full RAKE-receiver is used. The curves marked with triangles show the minimum value of *E*rx/*E*tx which occurs within these positions, again assuming a full RAKE. Thus the small-scale fading effect becomes visible. The curves with circles depict the normalized receive energy for a receiver which exploits only the strongest propagation path, i.e., a single correlator is applied. Transmitters-receiver separation is about 3 m, the carrier frequency is always set to 6.85 GHz.

Non-coherent UWB transmission is an attractive approach especially if simple and robust implementations with a small power consumption are required. Main application fields are low data rate sensor or personal area networks, which require low cost devices and a long battery life time. It should be noted that the current IEEE802.15.4a UWB-PHY for low data rate communications enables non-coherent detection, too [18]. The main advantage of a

**Figure 2.** Capacity of an AWGN channel as a function of the receive-power for different signal bandwidths. The 2.4 GHz ISM band offers a bandwidth of 83 MHz.

non-coherent receiver is clearly the dramatically reduced effort which is required for channel estimation, synchronization, and multipath diversity combining. This advantage is, however, bought by a serious drawback: non-coherent detectors are more susceptible to narrowband interference (NBI), multi-user interference (MUI), and inter-symbol interference (ISI).

Non-coherent detection can either rely on envelope detection or on differential detection. In the simplest case, path-diversity combining is carried out via an analog integration device. However, the change from analog to digital combining stimulates new perspectives. Since a digital code matched filter can be applied prior to the non-coherent part of the receiver, the capability to distinguish users (or networks) by means of code division multiple access is improved. We show that digital receiver implementations with user specific filtering have also an enhanced interference rejection capability and energy efficiency. Moreover, we present well suited solutions for the analog-to-digital conversion, the spread-spectrum code sequences, and the modulation format.
