Detecting bogus messages in vehicular ad-hoc networks: An information fusion approach

Abstract

In Vehicular ad hoc networks (VANETs), vehicles are allowed to broadcast messages for informing nearby vehicles about road condition and emergent events, such as traffic congestion or accident. It leaves a backdoor in which inside attackers can launch false information attacks by injecting bogus emergency messages to mislead other vehicles, and potential threats on road safety can be caused. This paper presents a multi-source information fusion approach to detect bogus emergency messages, in which each vehicle uses its on-board sensor data and received beacon messages to perceive the traffic condition and calculates its belief on credibility for received emergency messages. Moreover, the proposed approach provides enhanced robustness against collusion attacks by integrating an outlier detection mechanism in which a clustering algorithm is performed to filter out the colluder whose behavior deviates largely from others. The simulation results show validity of our approach, higher significantly detection rate can be achieved comparing to the existing threshold based scheme.