Enhancing Autonomous Vehicles System Security: Advanced Attack Detection for Robust Safeguarding

The advances in highly automated and autonomous transportation systems over the last decade have generated great interest in topics in the safe navigation of land vehicles. With distributed control strategies employed in the majority of applications of autonomous vehicles, such as traffic and formation control, the much-required resilience takes the form of fault-tolerance with respect to information corruption, especially, when such information is utilized in closed-loop control. This study addresses the topic of detection of malicious attacks in a decentralized traffic control system for land vehicles. The proposed method employs trajectory prediction based on a hierarchical Model Predictive Control scheme, as well as, vehicle-to-vehicle communication in order to generate redundancy of collision risk information. The efficacy of the method is demonstrated in Eclipse Simulation of Urban Mobility (SUMO) considering the scenario of junction traffic management.