Adaptive routing algorithm for vehicular networks in city environment

Abstract

The most known position based routing protocol for high mobile vehicular ad-hoc networks is the Greedy Perimeter Stateless Routing (GPSR) protocol. This protocol contains two routing modes, the greedy mode and the recovery mode, and creates a neighbor list to make a routing decision. The GPSR uses the greedy forwarding method whereas hops that follow are chosen based on nodes which are geographically closer to the destination node among the neighboring nodes. Nevertheless, forwarding fails if the current node is geographically the closest but unable to forward the packets to the destination. In this situation, the GPSR algorithm acts in recovery (perimeter) mode which is the repair strategy of this algorithm. Some enhancements such as the Greedy Perimeter Coordinator Routing (GPCR) and the Greedy Border Superiority Routing (GBSR) have been proposed. In this thesis, a routing protocol which works on real city maps and takes into account the actual movements of vehicles in city environment is recommended. We describe Adaptive Routing Algorithm For Vehicular Networks (ARAV) as a solution that improves the packet delivery ratio of GPSR and also improves the GPCR with road vehicle density information. ARAV differs from GPCR and GBSR, in that it uses routing algorithms adaptively in two routing modes. In ARAV greedy mode, we address the shortcomings of GPCR by considering the density of the paths. Packets are routed from one junction to another, the direction and position of hops are determined by taking the density of the paths into consideration. Our simulation results show that the proposed ARAV protocol outperforms the GPSR protocol in terms of packet delivery, throughput, delay in one successful transmission and greedy/perimeter ratio.