DSVL: Detecting Selfish Node in Vehicular Ad-hoc Networks (VANET) by Learning Automata

Abstract

Vehicular Ad-hoc Networks (VANETs) are a set of mobile nodes that move on the road and connect via wireless. Due to the limited radio range, they send data to each other by collaborating. Some nodes drop the other nodes’ packets to save the network supplements; therefore, the network’s performance will reduce. So it is necessary to identify selfish nodes to prevent other nodes from cooperating with them. In the proposed scheme, a punishmentbased algorithm is presented to identify the selfish nodes used in Adaptive Resonance Theory (ART) clustering to monitor and control them. The cluster head determines if selfish behaviors occur in the cluster or not. If the cluster head discovers that there is a selfish behavior in the cluster, it begins to check the packets that were sent and received by all nodes. In the proposed method, each node in the network is equipped with learning automata, the probability of selecting each neighbor node to send the packet, which is rewarded or punished according to the performance. Simulation results have shown that the rate of detection of selfish nodes is more than other methods, and the false alarm rate (FAR) is less than other similar methods.