Adaptive compliant skill learning for contact-rich manipulation with human in the loop

Si, Weiyong; Guan, Yuan; Wang, Ning

Abstract

It is essential for the robot manipulator to adapt to unexpected events and dynamic environments while executing the physical contact-rich tasks. Although a range of methods have been investigated to enhance the adaptability and generalization capability of robot manipulation, it is still difficult to perform complex contact-rich tasks, e.g., rolling pizza dough and robot-assisted medical scanning, without the assistance from a human in the loop. We proposed a novel framework combining learning from demonstration (LfD) and human experience to enhance the safety and adaptability of the robot manipulation. In this framework, dynamic movement primitives (DMPs) is employed for manipulation skills learning from demonstrations, and human correction is applied to update the pre-trained DMPs skills model. We conducted experiments on the Franka Emika Panda Robot with pizza dough rolling tasks. The results demonstrate that the proposed framework could effectively improve the performance of the physical contact-rich tasks, and the human correction method through teleoperation provides a potential solution for advanced interaction tasks with complex and dynamic physical properties.

• Journal Article Type: Article
• Acceptance Date: Jan 13, 2022
• Online Publication Date: Mar 15, 2022
• Publication Date: 2022-07
• Deposit Date: Apr 7, 2022
• Publicly Available Date: Apr 8, 2022
• Journal: IEEE Robotics and Automation Letters
• Print ISSN: 2377-3766
• Publisher: Institute of Electrical and Electronics Engineers
• Peer Reviewed: Peer Reviewed
• Volume: 7
• Issue: 3
• Pages: 5834 - 5841
• DOI: https://doi.org/10.1109/LRA.2022.3159163
• Keywords: Index Terms-Bilateral teleoperation; Compliant movement skills; Learning from demonstration; Force and impedance control
• Public URL: https://uwe-repository.worktribe.com/output/9303787