Memory-based event-triggered fuzzy path tracking security control for autonomous vehicles against stochastic false data injection attacks.

Controller area network (CAN)-dependent autonomous vehicles (AVs) control system is vulnerable to limited network bandwidth and stochastic false data injection (FDI) attacks, presenting a potential threat to the path tracking performance for AVs. To address this challenge, this article investigates a novel event-triggered robust security control method to enhance the system safety and efficiency. Firstly, an interval type-2 Takagi-Sugeno (IT2 T-S) fuzzy model is established to depict lateral vehicle dynamics. Secondly, an IT2 T-S fuzzy observer is designed to estimate the lateral speed of the vehicle. Then, a novel memory-based event-triggered mechanism (METM) is proposed to balance the network burden and control performance. After that, an observer-based fuzzy security controller is developed to guarantee H∞ stability by Lyapunov-Krasovskii functionals (LKFs) and the linear matrix inequalities (LMIs) approach. Finally, simulation and CAN-based semi-physical hardware-in-loop (HiL) results reveal that, compared with a typical H∞ output feedback controller, the proposed method can reduce the root mean square values of lateral error and yaw angle error by more than 31% and 11%, respectively.