Finite-Time Path-Following Control for an Autonomous Vehicle with Actuator Saturation Using Critic-Actor Network and Fractional-Order Sliding Mode Surface

This paper proposes a finite time path tracking control strategy for autonomous vehicle, which solves the challenges of actuator saturation, system uncertainty and external interference. This method combines the Critic Actor neural network framework with the sliding mode control method. We introduced the Integrated Fast Terminal Sliding Surface (IFTSMS) and designed a convergence law to achieve fast convergence to the sliding surface, ensuring stability and robustness under dynamic uncertainty. The Critic Actor neural network improves path tracking by optimizing control strategies and evaluating their effectiveness, while designing a compensation term to address actuator saturation issues. Comparative simulation experiments have demonstrated the effectiveness of the proposed method, which improves tracking performance and enhances robustness in the presence of persistent interference and actuator limitations. The analysis based on Lyapunov further ensures the stability of the system and verifies the finite time convergence and robustness of the proposed control method.