**5.1.1 CSMA/CA protocol**

Carrier Sense Multiple Access with Collision Detection protocol is a distributed control protocol which does not require any central coordinator. The principle of this protocol is that a transmitter that wants to initiate a transmission, checks the transmission channel by checking the presence of a carrier signal. If no carrier signal is present which indicates the channel is free and the transmitter can initiate a transmission. For a high propagation delay network such a solution does not offer very high throughput due to the delay.

Distance = d (m), Propagation delay = tp

Fig. 15. CSMA/CA protocol based packet transmission example

Consider Figure 15, where two nodes are using CSMA/CA protocol, are spaced apart by 100 meters. In this case, if at t=0, Node A senses the channel then it will find the channel to be free and can go ahead with the transmission. If Node A starts transmission of a packet immediately then it can assume that the packet will be successfully transmitted. However, if Node B starts sensing the channel before the propagation delay time *tp* then it will also find the channel is free and could start transmission. In this case both packet will collide and the transmission channel capacity will be wasted for a period of L+*tp* where L is the packet transmission time. On the other hand, if Node B checks the channel after time *tp* from the commencement of A's packet transmission, then it will find the channel is busy and will not transmit any packets. Now this simple example shows how the performance of random access protocol is dependent on the propagation delay. If propagation delay is small then there is much lower probability that a packet will be transmitted before the packet from A arrives at B. As the propagation delay increases the collision probability will also increase.

The CSMA/CA protocol is generally used in RF (Radio Frequency) networks where 100 m link delay will incur a propagation delay of 0.333 *μ*sec whereas an underwater acoustic link 22 Will-be-set-by-IN-TECH

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

Carrier Sense Multiple Access with Collision Detection protocol is a distributed control protocol which does not require any central coordinator. The principle of this protocol is that a transmitter that wants to initiate a transmission, checks the transmission channel by checking the presence of a carrier signal. If no carrier signal is present which indicates the channel is free and the transmitter can initiate a transmission. For a high propagation delay

Distance = d (m), Propagation delay = tp

Consider Figure 15, where two nodes are using CSMA/CA protocol, are spaced apart by 100 meters. In this case, if at t=0, Node A senses the channel then it will find the channel to be free and can go ahead with the transmission. If Node A starts transmission of a packet immediately then it can assume that the packet will be successfully transmitted. However, if Node B starts sensing the channel before the propagation delay time *tp* then it will also find the channel is free and could start transmission. In this case both packet will collide and the transmission channel capacity will be wasted for a period of L+*tp* where L is the packet transmission time. On the other hand, if Node B checks the channel after time *tp* from the commencement of A's packet transmission, then it will find the channel is busy and will not transmit any packets. Now this simple example shows how the performance of random access protocol is dependent on the propagation delay. If propagation delay is small then there is much lower probability that a packet will be transmitted before the packet from A arrives at

B. As the propagation delay increases the collision probability will also increase.

The CSMA/CA protocol is generally used in RF (Radio Frequency) networks where 100 m link delay will incur a propagation delay of 0.333 *μ*sec whereas an underwater acoustic link

**Node A Node B** 

CSMA/CA protocol and its applicability for underwater applications.

network such a solution does not offer very high throughput due to the delay.

Fig. 15. CSMA/CA protocol based packet transmission example

**5.1.1 CSMA/CA protocol**

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 throughput will be further reduced.

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 reduce the effective packet error rate.

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

$$\text{BER} = 1 - (1 - \text{BER})^K \tag{14}$$

#### **5.2 Packet routing**

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

Communication Networks 25

Short-Range Underwater Acoustic Communication Networks 197

Aldawibio, O. (2008). A review of current routing protocols for ad hoc underwater acoustic

Chen, W. & Mitra, U. (2007). Packet scheduling for multihopped underwater acoustic

Chirdchoo, N., Soh, W. & Chua, K. (2008). Ript: A receiver-initiated reservation-based protocol

Domingo, M. (2008). Overview of channel models for underwater wireless communication

Dunbabin, M., Roberts, J., Usher, K., Winstanley, G. & Corke, P. (2005). A hybrid auv

Eckart, C. (1952). *Principles of Underwater Sound*, Research Analysis Group, National Research

Essebbar, A., Loubet, G. & Vial, F. (1994). Underwater acoustic channel simulations for

Fisher, F. & Simmons, V. (1977). Sound absorption in sea water, *Journal of the Acoustical Society*

Francois, R. & Garrison, G. (1982). Sound absorption based on ocean measurements: Part 1 and 2, *Journal of the Acoustical Society of America* 72(3,6): 896–907, 1879 – 1890. Guangzhong, L. & Zhibin, L. (2010). Depth-based multi-hop routing protocol for underwater

Guo, X., Frater, M. & Ryan, M. (2009). Design of a propagation-delay-tolerant mac protocol for

Hajenko, T. & Benson, C. (2010). The high frequency underwater acoustic channel, *IEEE*

Holmes, J., Carey, W., Lynch, J., Newhall, A. & Kukulya, A. (2005). An autonomous

Johnson, M., Preisig, J., Freitag, L.& Stojanovic, M. (1999). FSK and PSK performance of the

Johnson, R. (2011). *University Corporation of Atmospheric Research, Window to the Universe*,

Karl, H. & Willig, A. (2006). *Protocols and Architecures for Wireless Sensor Networks*, John Wiley

Caruthers, J. (1977). *Fundamentals of Marine Acoustics*, Elsevier Scientific Publishing.

communication networks, *IEEE OCEANS'07*, pp. 1–6.

Coates, R. (1989). *Underwater Acoustic Systems*, John Wiley and Sons. Cox, A. W. (1974). *Sonar and Underwater Sound*, Lexington Books.

networks, *Physical Communication* pp. 163 – 182.

*Tomorrow's Preservation.'*, Vol. 3, pp. III/495 –III/500 vol.3. Etter, P. (2003). *Underwater Acoustic Modeling and SImulation*, third edn, Spon Press.

*IEEE Oceans 2005 - Europe*, Vol. 2, pp. 1058 – 1061 Vol. 2.

*Automation (ICRA)*, pp. 2117–2122.

Council, California University.

*(ICIMA)*, Vol. 2, pp. 268 –270.

*OCEANS 2010, Sydney*, pp. 1 –3.

tour/link/earth/Water/overview.html.

Vol. 3, pp. 1512 Vol. 3.

and Sons, Ltd.

*Web Technologies. ICADIWT*, pp. 431 –434.

networks, *First International Conference on the Applications of Digital Information and*

for underwater acoustic networks, *IEEE Journal on Selected Areas in Communications*

design for shallow water reef navigation, *Proc. International Conference on Robotics and*

communication, *IEEE OCEANS '94. 'Oceans Engineering for Today's Technology and*

sensor network, *2nd International Conference on Industrial Mechatronics and Automation*

underwater acoustic sensor networks, *IEEE Journal of Oceanic Engineering* 34(2): 170

underwater vehicle towed array for ocean acoustic measurements and inversions,

utility acoustic modem, *IEEE OCEANS '99 MTS. Riding the Crest into the 21st Century*,

University Corporation of Atmospheric Research, http://www.windows. ucar. edu/

**7. References**

26(9): 1744 –1753.

*of America* 62(3).

–180.

in a different manner. In this case all the link state information is periodically transmitted to all nodes in the network. In case of any change of state of a link, all nodes get notification and modify their routing table. In a swarm network link qualities will be variable which will require regular reconfiguration of routing tables. The performance of routing algorithms is generally determined by a number of factors including the convergence delay. In the case of a swarm network the convergence delay will be a critical factor because of high link delays. For underwater swarm applications, each update within a network will take considerably longer time than a RF network, causing additional packet transmission delays. Hence, it is necessary to develop the network structure in different ways than a conventional sensor network. For example, it may be necessary to develop smaller size clustered networks where cluster heads form a second tier network. Within this topology, local information will flow within the cluster and inter-cluster information will flow through the cluster head network. Cluster based communication architectures are also being used in Zigbee based and wireless personal communication networks (Karl & Willig, 2006). Further research is necessary to develop appropriate routing algorithms to minimise packet transmission delay in swarm networks. Readers can consult the following references to follow some of the recent progress in the area (Aldawibio, 2008; Guangzhong & Zhibin, 2010; LeonGarcia & Widjaja, 2004; Zorzi et al., 2008).

Discussion in this section clearly shows that the MAC and routing protocol designs require transmission channel state information in order to optimise their performance. Due to the high propagation delay of an underwater channel, any change of link quality such as SNR will significantly affect the performance of the network. Hence, it is necessary to develop a new class of protocols which can adapt themselves with the varying channel conditions and offer reasonable high throughput in swarm networks.
