Spatial Repetitive Learning Control for Trajectory Learning in Human-Robot Collaboration

Abstract

In this paper, a spatial repetitive learning control method is proposed for a robotic manipulator to learn a target trajectory that is designed by a human partner. It is different from the traditional teaching by demonstration. The proposed method updates the reference trajectory in a gradual learning manner that is based on spatial domain analysis, which addresses limits in the time-based learning in the existing literature. As the learning time period is not fixed, this approach can be used in more general tracking tasks. Stability of the control system is guaranteed by rigorous Lyapunov-stability analysis. In the end, simulation results show that the proposed spatial repetitive learning control method is effective.