Deep-Reinforcement-Learning-Based Active Disturbance Rejection Control for Lateral Path Following of Parafoil System

The path-following control of the parafoil system is essential for executing missions, such as accurate homing and delivery. In this paper, lateral studied. First, considering relative motion between canopy payload, an eight-degree-of-freedom (DOF) model constructed. Then, a guidance law containing position deviation heading angle proposed. Moreover, linear active disturbance rejection controller (LADRC) designed based on to allow track desired path under internal unmodeled dynamics or external environmental disturbances. For adaptive tuning parameters, deep Q-network (DQN) applied LADRC-based system, parameters can be adjusted in real time according system’s states. Finally, effectiveness proposed method following circular straight paths environment with wind simulation results show that effective means realize it also promote development intelligent controllers.