A VELOCITY SELF-LEARNING ALGORITHM FOR TIME-OPTIMAL TRAJECTORY PLANNING ALONG the FULLY SPECIFIED PATH – PART 1

In response to the problem of uncertainty in the system dynamics model during the time-optimal trajectory planning for the industrial handling robots, a novel online self-learning model-free time-optimal trajectory planning method is proposed. First, offline kinematic constraints and the Hermite interpolation algorithm are used to obtain the optimal spline velocity curve under kinematic constraints. Then, online trajectory data of the robot's operation is collected, and the trajectory generation method using a self-learning strategy is employed to refine the trajectory iteratively, resulting in a time-optimal trajectory under actual dynamic constraints. Finally, taking the ur5e cooperative robot and ABB IRB9670-235 industrial robot as the experimental platform, experiments verify the effectiveness and efficiency of the proposed method.