Collaborative trust aware intelligent intrusion detection in VANETs

Trust aware Collaborative Learning Automata based Intrusion Detection System (T-CLAIDS) for VANETs is proposed in this paper. Learning Automata (LA) are assumed to be deployed on vehicles in the network to capture the information about the different states of the vehicles on the road. A Markov Chain Model (MCM) is constructed for representation of states and their transitions in the network. Transitions from one state to other are dependent upon the density of the vehicles in a particular region. A new classifier is designed for detection of any malicious activity in the network and is tuned based upon the new parameter called as Collaborative Trust Index (CTI) so that it covers all possible types of attacks in the network. An algorithm for detection of abnormal events using the defined classifier is also proposed. The results obtained show that T-CLAIDS performs better than the other existing schemes with respect to parameters such as false alarm ratio, detection ratio and overhead generated.