Automatic Incident Detection In VANETs: A Bayesian Approach

Although vehicular ad hoc networks (VANETs) started mainly for safety applications, surprisingly very few work have been done in VANETs for automatic incident detection (AID) while most of the research went for developing routing protocols and privacy techniques. On the other hand, it is fundamentally difficult for most of the existing AID techniques to detect incidents in non dense traffic especially for those incidents that do not block all lanes. In this paper, we introduce a novel probabilistic automatic incident detection technique for non dense traffic flow based on Bayesian theory.