Resilient Predictive Control of Constrained Connected and Automated Vehicles under Malicious Attacks

In this paper, we present a novel resilient distributed model predictive control (RDMPC) framework for the constrained Connected and Automated Vehicles (CAV) in the presence of F-local malicious attacks. The proposed framework aims to ensure constraint satisfaction and identify malicious attacks using previously broadcast information and a convex set, referred to as the "resilience set." Compared to the well-known Mean Subsequence Reduced (MSR) algorithms that require (2F + 1)robust graphs, the proposed approach significantly reduces the required robustness level to (F + 1)-robust graph. Our simulation results demonstrate the effectiveness of the proposed approach in mitigating the impact of malicious attacks on constrained CAVs while ensuring constraint satisfaction. Overall, the proposed RDMPC framework contributes to the field of resilient platoon control for CAVs and has potential implications for improving the reliability and security of CAVs in real-world scenarios.