Highly efficient approach for discordant BSMs detection in connected vehicles environment

Connected Vehicles (CVs) are the key enabling technology for Intelligent Transportation Systems (ITSs) that offer great opportunities for improving traffic safety and efficiency. They provide several innovative safety-related applications such as traffic management and monitoring, which involve the transmission of messages from all vehicles on the road. Basic Safety Messages (BSMs) constitute an essential type of control message. However, several critical issues affect the BSM messages’ reliability. In this paper, a model-based approach for detecting discordant BSMs is proposed, which allows to avoid the vehicle disturbance. This approach consists of detecting incoherence in communication metric values, where the detection is formulated as an anomaly detection problem that is solved using the Gaussian distribution. The detection process allows the vehicles to cross their prediction to achieve more precision in deciding whether to accept or reject a message from a vehicle. The efficiency of our model for detecting an anomaly has been evaluated through simulations using our generated dataset. The obtained results indicate that the proposed model provides high performance in terms of detection rate. Moreover, we evaluate and validate the proposed approach through formal evaluation, where it demonstrates promising performances, as compare it with a concurrent approach through simulations considering important metrics.