Force-dependent variable impedance controller for contact-rich tasks under reference trajectory uncertainty

Abstract

In robotic manipulation, performing force-tracking tasks in an uncertain environment poses the risk of the robot and the environment encountering high contact forces. While learning control methods are used wheninteracting with uncertain environments, the robots generally take some time to learn the correct reference path in such scenarios. During this process, it is important to reduce the contact forces until the environment properties are learned to ensure the safety of the interaction. To this end, this article proposes a force-dependent variable impedance controller (FVIC) that provides compliance in the presence of reference uncertainty and improves the position tracking accuracy as the certainty of the reference position increases. In this FVIC, the stiffness and dampingof the robot are defined as functions of force and force rate, respectively, to ensure compliance and stability. The proposed method is validated via simulations and experiments conducted using the Kinova Gen3 7DOF robot. The results show that, unlike the traditional variable impedance control (VIC) methods, this method ensures stability without compromising the desired impedance characteristics. It is further demonstrated that with this method, the contact forces can be maintained significantly low when there’s a reference uncertainty, thus ensuring safety.

More information

Identifiers

ORCID for Hareesh Godaba: orcid.org/0000-0001-6600-8513

Catalogue record

Export record

Altmetrics