IDENTIFICATION OF MALICIOUS BEHAVIOUR OF VEHICLE ON VANET USING SOM CLASSIFIER

Abstract: The vehicular ad-hoc network now a day's growing field of research, due its infrastructure or rapidly change topology. VANET is sub part of MANET and combination of nodes and roadside units. VANET uses high movable nodes as compared to MANET. VANET provide wireless communication among vehicles and vehicle to roadside unit for sharing information and safety purpose of drivers and passengers. There are various malicious activities performed in network like bogus information attack, ID discloser, sybil attack etc. All these attacks try to distract drivers. In this paper we work on Dos attack in AODV routing protocol. When malicious node sends fake requests frequently to other nodes it creates a blockage in network then node is not able to respond to other nodes. In this paper Artificial Neural Network in VANET is used; so neural network helps to train the node and uses the back propagation and adjust the weights. For the identification of malicious node SOM classifier is used. SOM observe the behavior of nodes and classifies as the normal node and malicious node in the network.

Keywords: VANET, ANN, DOS Attack, AODV, security, SOM classifier.