Redundancy Resolution in Human-Robot Co-manipulation with Cartesian Impedance Control

Abstract

In this paper the role of redundancy in Cartesian impedance control of a robotic arm for the execution of tasks in co-manipulation with humans is considered. In particular, the problem of stability is experimentally investigated. When a human operator guides the robot through direct physical interaction, it is desirable to have a compliant behaviour at the end effector according to a decoupled impedance dynamics. In order to achieve a desired impedance behaviour, the robot’s dynamics has to be suitably reshaped by the controller. Moreover, the stability of the coupled human-robot system should be guaranteed for any value of the impedance parameters within a prescribed region. If the robot is kinematically or functionally redundant, also the redundant degrees of freedom can be used to modify the robot dynamics. Through an extensive experimental study on a 7-DOF KUKA LWR4 arm, we compare two different strategies to solve redundancy and we show that, when redundancy is exploited to ensure a decoupled apparent inertia at the end effector, the stability region in the parameter space becomes larger. Thus, better performance can be achieved by using, e.g., variable impedance control laws tuned to human intentions.

This research has been partially funded by the EC Seventh Framework Programme within the SAPHARI project 287513 and RoDyMan project 320992.