**5. Swarm network protocol design techniques**

20 Will-be-set-by-IN-TECH

signalling with strong error correction coding that provides some resilience to the rapidly varying multipath. Alternatively, the use of a higher rate coherent method of QPSK signalling that incorporates a Doppler tolerant multi-channel adaptive equalizer has gained in appeal

The BER formulae are well known for FSK and QPSK modulation techniques (Rappaport,

= *SNR*(*r*) ×

where *Rb* is the data rate in bps and *Bc* is the channel bandwidth. Equation 12 and 13 are the

2 *er f c*[ *Eb No* ]

*No* (b) BER vs Range vs *Eb*

Fig. 13. Probability of Bit Error for Short Range Acoustic Data Transmission Underwater

The data rates *Rb* used are 10 and 20 kbps to reflect the current maximum commercial

power of 10mW and a data rate of 20kbps. This increases to 12dB if using FSK with half the data rate (10 kbps) and same Transmitter Power. From Figure 13 (b), these settings will provide only a 150 m range. The range can be increased to 250 m using QPSK if the data rate was halved to 10 kbps or out to 500 m if the transmitter power was increased to 100mW in addition to the reduced data rate. Transmitter power plays a critical role, as illustrated here, by the comparison of ranges achieved from ≈ 75 m to 500 m with a change of transmitter

2 *er f c*[ 1 2 *Eb No* ]

*Eb No*

*QPSK* : BER <sup>=</sup> <sup>1</sup>

*FSK* : BER <sup>=</sup> <sup>1</sup>

*No*

*Bc Rb* , that can be found from the SNR

1/2 (12)

1/2 (13)

*No* (for QPSK)

*No* and Range respectively. Taking a

*No* required for QPSK is 8dB for a transmitter

(11)

over that time (Johnson et al., 1999).

(Equation 8) by:

1996), which require the Energy per Bit to Noise psd, *Eb*

uncoded BER for BPSK/QPSK and FSK respectively:

(a) BER vs *Eb*

achievable levels. Figure 13 (a) and (b) show the BER for *Eb*

BER of 10−<sup>4</sup> or 1 bit error in every 10, 000 bits, the *Eb*

power needed from 1mW to 100mW for this BER.

A short range underwater network, as shown in Figure 1(b) is essentially a multi-node sensor network. To develop a functional sensor network it is necessary to design a number of protocols which includes MAC, DLC (Data Link Control) and routing protocols. A typical protocol stack of a sensor network is presented in Figure 14. The lowest layer is the physical layer which is responsible for implementing all electrical/acoustic signal conditioning techniques such as amplifications, signal detection, modulation and demodulation, signal conversions, etc.. The second layer is the data link layer which accommodates the MAC and DLC protocols. The MAC is an important component of a sensor networks protocol stack, as it allows interference free transmission of information in a shared channel. The DLC protocol includes the ARQ (Automatic Repeat reQuest) and flow control functionalities necessary for error free data transmission in a non zero BER transmission environment. Design of the DLC functionalities are very closely linked to the transmission channel conditions. The network layers main operational control is the routing protocol; responsible for directing packets from the source to the destination over a multi-hop network. Routing protocols keep state information of all links to direct packets through high SNR links in order to minimise the end to end packet delay. The transport layer is responsible for end to end error control procedures which replicates the DLC functions but on an end to end basis rather than hop to hop basis as implemented by the DLL. The transport layer could use standard protocols such as TCP (Transmission Control Protocol) or UDP (User Datagram Protocol). The application layer hosts different operational applications which either transmit or receive data using the lower layers. To develop efficient network architectures, it is necessary to develop network and/or application specific DLL and network layers. The following subsections will present MAC and routing protocol design characteristics required for underwater swarm networking.

Fig. 14. A typical protocol stack for a sensor network

#### **5.1 MAC protocol**

Medium access protocols are used to coordinate the transmission of information from multiple transmitters using a shared communication channel. MAC protocols are designed to maximise channel usage by exploiting the key properties of transmission channels. MAC protocols can be designed to allocate transmission resources either in a fixed or in a dynamic manner. Fixed channel allocation techniques such as Frequency Division Multiplexing (FDM) or Time Division Multiplexing (TDM) are commonly used in many communication systems where ample channel capacity is available to transmit information (Karl & Willig, 2006). For low data rate and variable channel conditions, dynamic channel allocation techniques

Communication Networks 23

Short-Range Underwater Acoustic Communication Networks 195

of same distance will generate a propagation delay of 0.29 sec which is about 875,000 times longer than the RF delay. One can easily see why an acoustic link will produce much lower throughput than is predicted by the Shannon-Hartley theorem as discussed in Section 4.3. If we assume that we are transmitting a 100 byte packet, then the packet will take about 0.08 sec to transmit on a 10 kbps RF link. The same packet will take 0.3713 sec on a 10 kbps acoustic link offering a net throughput of 2.154 kbps. This calculation is based on the assumption that the transmission channel is ideal i.e. BER=0. If the BER of the channel is non zero then the

Previous sections have shown that the BER of a transmission link is dependent on the link parameters, geometry of the application environment, modulation techniques, and presence of various noise sources. Non zero BER conditions introduce a finite packet error rate (PER) on a link which is described by Equation 14, where K represents the packet length. The PER will depend on the BER and the length of the transmitted packet. For a BER of 10−<sup>3</sup> using a packet size of 100 bytes, the link will generate a PER value of 0.55 which means that almost every second packet will be corrupted and require some sort of error protection scheme to

There are generally two types of packet error correction techniques used in communication systems, one is forward error correction (FEC) scheme which uses a number of redundant bits added with information bits to offer some degree of protection against the channel error. The second technique involves the use of packet retransmission techniques using the DLC function known as the ARQ. The ARQ protocol will introduce retransmissions when a receiver is unable to correct a packet using the FEC bits. The retransmission procedure could effectively reduce the throughput of a link further because the same information is transmitted multiple times. From this brief discussion one can see that standard CSMA/CA protocols used in sensor networks are almost unworkable in the underwater networking environment unless the standard protocol is further enhanced. This is a major research issue which is currently followed up by many researchers and authors. Readers can find some of the current research work on the MAC protocol in the following references (Chirdchoo et al., 2008; Guo et al., 2009;

Packet routing is another challenging task in the underwater networking environment. Packet routing protocols are very important for a multi-hop network because the receivers and the transmitters are distributed in a geographical area where nodes can also change their positions over time. Each node maintains a routing table to forward packets through multi-hop links. Routing tables are created by selecting the best cost paths from transmitters to receivers. The cost of a path can be expressed in terms of delay, packet loss, BER, real monetary cost \$, etc.. For underwater networks, the link delay could be used as a cost metric, to transmit packets with a minimum delay. Routing protocols are generally classified into two classes: distance vector and link state routing protocols (LeonGarcia & Widjaja, 2004). The distance vector algorithms generally select a path from a transmitter to receiver based on shortest path through neighbouring networks. When the status of a link changes, for example, if the delay or SNR of a link is increased then the node next to the link will detect and inform its neighbour about the change and suggest a new link. This process will continue until all the nodes in the network have updated their routing table. The link state routing protocols work

PER <sup>=</sup> <sup>1</sup> <sup>−</sup> (<sup>1</sup> <sup>−</sup> BER)*<sup>K</sup>* (14)

throughput will be further reduced.

reduce the effective packet error rate.

Pompili & Akyildiz, 2009; Syed et al., 2008).

**5.2 Packet routing**

are generally used to maximise the transmission channel utilisation where the physical transmission channel condition could be highly variable. Based on the dynamic channel allocation technique it is possible to develop two classes of MAC protocols known as random access and scheduled access protocol. The most commonly used random access protocols is the CSMA (Carrier Sense Multiple Access) widely used in many networks including sensor network designs. Most commonly used scheduled access protocol is the polling protocol. Both the CSMA and polling protocols have flexible structures which can be adopted for different application environments. As discussed in this chapter, the underwater communication channel is a relatively difficult transmission medium due to the variability of link quality depending on location and applications. Also, the use of an acoustic signal as a carrier will generate a significant delay which is a major challenge when developing a MAC protocol. In the following subsection we discuss the basic design characteristics of the standard CSMA/CA protocol and its applicability for underwater applications.
