Dynamic Self-Triggered Intelligent Path Tracking Control for Autonomous Agricultural Vehicles via Reinforcement Learning

This paper investigates the path tracking control of autonomous agricultural vehicles with disturbances under a dynamic self-triggered mechanism via reinforcement learning (RL). To begin with, a path tracking offset system is constructed based on the kinematic model of the autonomous agricultural vehicle, which transforms the path tracking control problem into an optimal control problem. Subsequently, a novel dynamic self-triggered second-order integral sliding mode control policy is developed to mitigate the impact of disturbances and to derive a nominal path tracking offset model. Afterward, to further alleviate the computing and communication burdens, a novel dynamic self-triggered mechanism is proposed for the optimal control policy. It can predict the next update time based on current information, thus avoiding the need for continuous monitoring of the triggering condition. Furthermore, a single critic network architecture is constructed to obtain an approximate path tracking control policy, and it is proven by Lyapunov stability theory that this policy ensures autonomous agricultural vehicles can maintain the predefined working path in the presence of disturbances. Finally, the effectiveness of the proposed path tracking control method is demonstrated by simulation experiments. Note to Practitioners—Autonomous agricultural vehicles serve as the core equipment in autonomous farms, and path tracking control is a critical technology for achieving precise and efficient operations. However, existing path tracking control methods for autonomous agricultural vehicles primarily focus on system stability, without considering the control cost. Moreover, traditional methods are typically time-triggered, leading to substantial consumption of computing and communication resources. To overcome the above bottleneck, this paper proposes an RL-based dynamic self-triggered path tracking control method for autonomous agricultural vehicles with disturbances. The proposed method comprises two components, that is, a second-order integral sliding mode control policy to mitigate the impact of disturbances and an optimal control policy to ensure that the autonomous agricultural vehicle follows the desired path. Notably, both control policies are implemented within the self-triggered mechanism, which predicts the next update time based on current information, such that the need for continuous state monitoring is eliminated. Simulation results demonstrate that the proposed method enhances the control performance and enables the completion of operational tasks in an efficient and energy-saving manner.