Cyber-Physical System-Based Path Tracking Control of Autonomous Vehicles Under Cyber-Attacks

As one of the typical applications of cyberphysical system (CPS), autonomous vehicles (AVs) are vulnerable to malicious disturbance from cyberattacks while tracking the desired path. This article focuses on CPS-based path tracking control problem of AVs under cyberattacks. First, the nonlinear state and measurement equations of AVs under cyberattacks are established based on vehicle dynamics model. Second, to improve the robustness of AVs against cyberattacks, sensor redundancy is introduced. A cyberattack detection method is designed by using extended Kalman filter, and the computational complexity of the cyberattack detection method is analyzed. In addition, sensor switching rules are developed to isolate the disturbance of cyberattacks. Then, the control problem of the AVs is formulated based on model predictive control. Input-to-state stability of the control system under cyberattacks is established, and a link between the tolerable attack intensity and the detection thresholds is clearly revealed. Finally, the simulation results demonstrate the effectiveness of the proposed control strategy.