Wireless Ad Hoc and Sensor Networks

Optimized Energy and Delay-Based Routing 387from other nodes in the network. RSSI is a widely used parameter forwireless networks in general. It is important to notice that the RSSI providesan indirect assessment of the channel state, which has to be takeninto account both during self-organization and routing. Channel state istotally ignored in the literature in most protocols. Once the CHs areselected, these CHs broadcast a beacon to all the nodes in the subnetwork.The nodes in the subnetwork measure the RSSI for these beacons, anddepending on the strength of each signal, the nodes join a CH with highestsignal strength. Once a node selects a CH, it sends a JOIN packet to theCH to indicate that it is joining a particular cluster. Sensor nodes in eachcluster will now relay information directly to the corresponding CHs. Bydoing this, they avoid flooding the network with data and allow the CHsto perform data aggregation so that the redundant data are compressed.After the network self-organizes itself, then the data is routed throughthe network as follows.8.8 Optimized Energy-Delay Subnetwork Routing ProtocolOnce the CHs are identified and the nodes are clustered within the subnetwork,the CHs initiate routing towards the BS by checking if the BS iswithin the communication range so that the data can be sent directly.Otherwise, the data from the CHs in the subnetwork have to be sent overa multihop route to the BS. The proposed on-demand routing algorithmis fully distributed and adaptive because it requires local information forbuilding routes and adapting to the topological changes. A high-powerbeacon from the BS is sent periodically to the entire network such that allthe nodes in the network have knowledge of the distance to the BS. It isimportant to note that the BS has a sufficient power source and thereforeit is not energy constrained. Though the OEDSR protocol borrows theidea of using an energy-delay type metric from the OEDR, the selectionof RNs is not based on maximizing the number of two-hop neighbors andavailable energy. Instead, the link cost factor is calculated by using a totallydifferent metric as given in Equation 8.4. Here, the selection of a relaynode depends on maximizing the link_cost_ factor, which includes thedistance from the relay node to the BS. Moreover, the route selection isperformed differently between the OEDR and the OEDSR because theOEDR is a proactive routing protocol whereas the OEDSR is an ondemandrouting protocol. Within the subnetworks, CHs are involved inthe routing process, whereas RNs are selected outside the subnetworksfor the purpose of routing. These RNs are viewed as CHs. Next, theselection of optimal RNs is discussed.