A Trajectory Planning Method Based on Imitation Learning for Hydraulic Robotic Manipulators
Xuli Liu, Yukun Zheng, Song Gao, Rui Qi Song, Rui Gao, Yibin Li · 2024
For hydraulic robotic manipulators trajectory planning, we propose an imitation learning method that combines Dynamic System (DS) models and probabilistic learning. Trajectory teaching is carried out based on the teleoperation system built in this article, and the trajectory is modeled, learned and reproduced based on the Dynamic System (DS) model and Task-parameterized Gaussian Mixture Model (TP-GMM) and Gaussian Mixture Regression (GMR). To validate this method, we conducted experiments where a 6-degree-of-freedom (6-DOF) hydraulic robotic manipulators wrote letters. The results demonstrate that this approach can successfully reproduce the trajectories of letters and generalize the starting and ending points of these trajectories.