Q-Learning-Based Trajectory Tracking Control of Delta Robot

This work investigates the trajectory tracking problem of the 3-degree-of-freedom Delta robot in the presence of unknown external disturbances, where the Q-learning-based reinforcement learning algorithm is utilized to optimize the PID controller. Firstly, the inverse solution of the Delta robot is established to determine the three rotation angles based on the pre-set motion coordinates. Then, the forward kinematics solution is constructed to convert the rotation angles back to the motion coordinates. The mathematical model of the Delta robot via the Jacobi matrix is further given. Furthermore, due to unknown disturbances, Q-learning is utilized to optimize the PID parameters in order to minimize the tracking errors. Finally, the simulation result is given to verify the effectiveness of the proposed method.