VANET Performance Improvement by Range Extending Agent and Stable Clustering

Abstract

Vehicular ad-hoc networks (VANETs) technology has emerged as an important research area over the last few years it is part of Mobile Ad-hoc Networks (MANET) that facilitates communication between groups of vehicles and roadside units (RSU) to provide safety and data communication. A challenging task in VANET is routing the message or disseminating the data to their destination due to high mobility and dynamism. Geomulticast is a specialized location-dependent multicasting technique, where messages are multicast to some specific user groups within a specific geographical zone. Enhanced Bandwidth Efficient Cluster Based Multicasting Protocol (EBECM) that improve the performance of VANET uses the concept of clustering for providing V2V communication and to be aware (ordinary vehicles) of the emergency situation by clustering multicast member with an ordinary vehicle. Stable cluster formation becomes difficult as it becomes tedious to cover all the vehicles in the clusters. On the other hand, non-clustered nodes are unable to communicate with cluster members and multicast nodes. In this case, it may lead to a multi-car chain collision. This study aims to improve VANET performance by extending the communication range using a special extend agent which enables RSU edge vehicles to provide backbone communication for out-of-range vehicles and be aware of accurate location and movement direction using GPS. Clusters formed based on speed, density, and direction of vehicles, inside RSU for multicast groups and outside the RSU range for non-multicast groups. Position and movement data gathered from vehicles information in communication range and by GPS from RSU out of communication range nearby one kilometer vehicles to aware physical and logical topology, to enhance member prediction and cluster stability. An extend agent is a vehicle that exists at the edge of the communication range and the nearest out of range vehicle to create continuance communication. The extend agents forward packets between multicast and non-multicast groups by layer 2 switching rather than layer 3 routing to reduce the routing overhead and minimize end-to-end delay. If an extend agent node (edge in RSU range vehicle) goes outside the RSU range, the agent module predicts the occurrence of link failure before the link failure extend agent module sets/selects another edge node as the extend agent node. Simulation results tested on ns2.35 with SUMO-1.15.0 to show the improved performance of the new proposed solution algorithm in terms of packet delivery ratio, routing overhead, and end-to-end delay when compared with EBECM protocol.