Modularity based mobility aware community detection algorithm for broadcast storm mitigation in VANETs

•An improved community detection algorithm based on local neighbourhood information.•The similarity between vehicles is computed based on Modularity Gain and Degree ofCohesion.•The application of Degree of Cohesion yields better stability in community formation.•The store-carry-forward mechanism addresses the intermittent connectivity for improving reachability in disconnected networks. High rates of road accidents have turned VANET to be a dynamic area of research, as it possibly enhances vehicle & road safety, traffic efficiency, and convenience. For such safety applications, VANET requires fast dissemination of messages to the nearby vehicles without any delay by rebroadcasting, which may lead to broadcast storm problem. To address this problem, the community detection algorithm is an efficient way of grouping similar vehicles into communities and selecting a few amongst them as forwarders. The existing research works forms the communities by considering the parameters like edge-betweenness, modularity gain, link stability, and neighbourhood similarity. Due to frequent switching between the communities these techniques suffer from ping-pong effect. Hence, a novel algorithm has been proposed which is based on modularity gain and degree of cohesion between vehicles to form stable communities. The proposed algorithm is unique amongst the state-of-the-art algorithms by selecting next forwarders of the safety messages by forming stable communities of vehicles considering their relative mobility. The simulation results confirm that the proposed algorithm forms stable communities and identifies less number of forwarders with reduced overhead and delay without reducing the percentage of vehicles that receives the message.