Kalman filter for trust estimation in VANETs

In this paper, we propose an automatic trust prediction mechanism based on Kalman filter for trust computation in VANET. The method fuses existing trust with the verified behavior of nodes to obtain an updated trust value iteratively. VANET is an ad hoc network characterized by high node mobility and dynamic topology, where the vehicles exchange messages among them to enhance driving conditions. Due to stringent time constraints imposed by the need of small time to react in case of safety conditions, message verification or node authentication is not possible. Trust evolution over a period of time is also precluded due to high node mobility and dynamic topology of VANET. Kalman filter can be used to quickly establish trust value and iteratively update it with the arrival of information. Results and simulations show that the predicted trust values are very close to the actual trust values and Kalman filter can easily capture behavior of both node with constant trust value and time varying trust values within a very small time span.