Density Based Geographic Routing Protocols in Vehicular Ad hoc Networks

Abstract

Vehicular ad hoc networks (VANETs) have become practical and valuable for their newlinewide variety of novel applications, such as road safety, content sharing, payment of newlinedifferent commercial places, etc. Due to high mobility and frequent disconnections of newlinenetwork established by the vehicles, geographic routing protocols that, do not require newlineroute formation prior to forwarding the actual data packets, are commonly adopted for newlineVANETs. Further, they also incur less overhead for route maintenance. Most of the newlineposition based routing protocols adopt the greedy mode and recovery mechanism to newlineestablish the connectivity. However, the presence of a void region can severely impact newlinethe performance of these protocols. Moreover, in case of highway as an environment, newlinethe existence of sparse and void regions tends to increase and the impact of these on newlinethe protocol can be alarming. Further, if the probability of perimeter mode (void newlineregion) is higher than the greedy mode, the end-to-end delay will increase, due to the newlinetime required to get the decision making strategy back from perimeter mode to greedy newlinemode. This thesis proposes an approach where a vehicle looking for next forwarder newlineselects the vehicle having sufficient density of immediate neighbours, and having newlineleast standard deviation of the average relative velocity in its neighbourhood. newlineHowever, the selected next best fit forwarder may or may not be the farthest from the newlinepacket holding vehicle and nearer to the destination. As the longevity and newlineassociativity of the connections can be observed to be dependent on the proposed newlinecriterion of next hop forwarder selection, the simulation results reflect a better newlineperformance as compared to the existing position based routing methods. newlineThe longevity and associativity of the connections can also be improved by providing newlinea structure with more stable topology in a highly dynamic and challenging topology newlinecharacteristic of VANETs. Clustering is one of the ways to do it. So, to maintain the newlinelongevity and associativity, the thesis proposes a dens