Learning Parameters of Damping Field using Bayesian Optimization for Human-robot Cooperation

Abstract

In impedance control, impedance parameters can be adjusted to improve the performance of human-robot collaboration. This paper proposes a new method to adapt the damping parameter based on the potential field method widely used in path planning. First, a damping field is generated to modify the damping value based on the current end-effector position. Bayesian optimization is then used to learn the parameters of the constructed damping field. Finally, experiments are conducted with a 2-DOF planar robot arm in a point-to-point collaborative motion, and results show that the proposed method help improve the performance of tasks that require high accuracy.