Dealing With Unexpected Moving Obstacles By Integrating Potential Field Planning With Inverse Dynamics Control

Abstract

We present a motion planning/control system that deals with moving obstacles whose trajectories are not known a priori. An artificial potential field planner is tightly coupled with a robust inverse dynamic controller, allowing the robot to avoid obstacles in realtime while retaining the benefits of inverse dynamic control. Our implementation of the artificial potential field planner uses digital filtering techniques to shape the input signal to the inverse dynamic controller, so that system’s response to moving obstacles will depend, not only on the position of those obstacles, but also on their velocity relative to the robot. We prove stability along solution trajectories of the system in the absence of obstacles, and discuss stability issues that arise when obstacles are present.