Elasto-Plastic Robot Compliance in Human-Robot Interaction and Robot-Robot Cooperation

Abstract

Efficient Human-Robot Interaction requires flexible role switching and robust role distribution. Leader and follower agents are mostly recognized based on interaction forces. This often comes with the drawback of reduced robustness and the necessity of models of the robot's environment or human behaviour. This video presents the recently proposed Elasto-Plastic Robot Compliance (EPRC) concept that, being power-based, additionally considers the robot velocity in role recognition. Thus, active environments can be detected inducing enhanced virtual plastic compliance of the robot. Such plastic compliance refers to evasive robot motions induced specifically by active environments, after which the robot is not pushed back to the initial point of contact in contrast to elastic compliance. The video further presents advantages as role distribution specific to different degrees-of-freedom, independence from model generation and unrestricted force application.