Disagreement-Aware Variable Impedance Control for Online Learning of Physical Human-Robot Cooperation Tasks
L.F. van der Spaa (Honda Research Institute Europe GmbH, TU Delft - Learning & Autonomous Control)
G. Franzese (TU Delft - Learning & Autonomous Control)
J. Kober (TU Delft - Learning & Autonomous Control)
Michael Gienger (Honda Research Institute Europe GmbH)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
In order to make the coexistence between humans and robots a reality, we must understand how they may cooperate more effectively. Modern robots, empowered with reliable controls and advanced machine learning reasoning can face this challenge. In this article, we presented a Disagreement- Aware Variable Impedance (DAVI) Controller, where the robot stiffness is regulated as a function of the perceived disagreement with the human cooperator. We tested the algorithm on a 7 DoF Franka Emika Panda robot performing the learning of a pick&place task with continuous adaptation of the goal location and the via-points with human interactive corrections, triggered by our proposed approach. A validation study was conducted with 5 users in order to understand the reliability of the method.