Minimally-Back-Drivable Robots for Rehabilitation: Path-Adherent Permissiveness Control via Trajectory Adaptation

Abstract

In pursuit of effective robot-assisted rehabilitation, it is imperative that the robot facilitates rather than hinders the patient's self-movements in appropriate directions. This essential attribute, termed permissiveness, is often lacking in conventional industrial robots. In this letter, an innovative approach is introduced, enabling the robot to provide a controlled degree of permissiveness in specified directions. Central to the method is a novel mapping function that dynamically adjusts the desired trajectory along the therapeutic path in response to the patient's monitored physical effects. The proposed approach is agnostic to the type of controlled robot employed, rendering even minimally-back-drivable robots viable for rehabilitation purposes. The letter outlines the approach's diverse facets and potential applications, including an iterative phase-based support adaptation policy, and a series of experiments showcasing the practical feasibility.