REINFORCEMENT LEARNING-BASED ADAPTIVE MOTION CONTROL FOR AUTONOMOUS VEHICLES VIA ACTOR-CRITIC STRUCTURE

In this paper, an optimized adaptive human-machine collaborative torque control methodology is developed for a class of autonomous vehicle systems with unknown disturbances by employing a reinforcement learning technology with identifier-critic-actor architecture. The criticism method is proposed to assess the performance of the control and to give feedback to the actor. Furthermore, a human-vehicle-road model with steering torque as input is constructed to describe the vehicle's dynamic response and the driver's maneuvering behavior. The stability of the control system is analyzed by means of the Lyapunov theory. Finally, the effectiveness of the control methodology is illustrated by means of a numerical example.