Anti-attack Trust Evaluation Algorithm Based on Bayesian Inference in VANET

Vehicular Ad-hoc Networks (VANETs) are crucial for intelligent transportation, improving traffic efficiency and safety. To enhance the security of VANET, trust management mechanism is implemented to defend against internal attacks in VANET. However, attacks targeting trust management mechanism, such as on-off attack, compromise trust management accuracy. In this paper, we propose the Anti-Attack Trust Evaluation Algorithm (AATEA) based on Bayesian inference to calculate trust values and establish reliable relationships among vehicles. AATEA addresses the challenge of on-off attack, where trust values are accumulated during continuous cooperation and suddenly initiate malicious behavior. Bayesian inference is employed to compute the trust values based on historical interactions. Additionally, we introduce an adaptive decay factor that considers the rate of change in trust values between the current and previous interaction of vehicles, to mitigate on-off attack. A dynamic driving reference set is designed based on the location information of received messages, since the forward and lateral vehicles of driving direction can provide more valuable information. Moreover, we built a VANET simulation platform using NS3 and SUMO, integrating security components, communication modules based on C-V2X, on-off attack and sybil attack module. Experimental results and comparisons with other VANET trust evaluation algorithms demonstrate AATEA’s superior performance in trust value principles.