Event-Triggered Output Feedback Model Predictive Control for Path Following of Autonomous Vehicles Under False Data Injection Attacks

Qing Sun, Xin Qi, Dajun Du, Xiaoqiang Ren, Minrui Fei

Abstract: Attacks on the sensor-controller (S-C) channel, which enable the infusion of false data, present substantial risks to the security of autonomous vehicles. This paper endeavors to address the challenge of guaranteeing secure path-following control for autonomous vehicles under the threat of False Data Injection (FDI) attacks, which through the utilization of an event-triggered output feedback model predictive control approach. To ensure the controller receives uncompromised system states amidst FDI attacks, a distributed robust multivariate observer (DRMO) is employed, which can facilitate the estimation and differentiation of uncompromised system states and attack signal simultaneously. Building upon the uncompromised system states provided by the observer and considering various constraint conditions, a predictive controller is formulated to ensure secure control of autonomous vehicle path following. Additionally, an event-triggered mechanism is introduced, dynamically adjusting the controller update frequency, resulting in significant savings in computational and communication resources. Finally, showcase an example to substantiate the efficiency of the proposed scheme.

External IDs:doi:10.1109/tiv.2024.3412428