A Bayesian Filter to Detect Misbehaving Nodes in MANETs

Composed of mobile nodes, MANETs rely on the entire collaboration of nodes to provide routing and forwarding functions. Most MANET security solutions identify and exclude malicious and selfish nodes. These mechanisms rely on local monitoring tools to detect untrustworthy nodes and then spread trust information through recommendations. In this paper we apply a Bayesian model [1] to improve MANET security. We classify nodes behavior relying on a decision based on a probabilistic Bayesian model. The novelty of our approach is that it classifies nodes based on distant observations. We describe statistical parameters like Accuracy, Error and the Total Cost Ratio that are used to evaluate the efficiency of the filter through a set of simulations of an attacked pure ad hoc network.