Synthesis and control of an assistive robotic tennis trainer

Abstract

While robotic guidance can improve ongoing performance of a motor task, several studies have found that it ultimately impairs motor learning. However, some recent studies found that the haptic demonstration of optimal timing, rather than movement magnitude, apparently transfers to participants trained with robotic guidance. Here, we describe the synthesis, design and control of a novel tendon-based haptic device able to apply different forms of physical guidance to train a task for which timing is the critical performance determinant: a forearm tennis stroke. We designed three guidance control strategies: a position controller to apply a fixed haptic guidance, a path controller with zero-force controller inside a safety tunnel around the desired trajectory to apply no guidance, and a guidance-as-needed controller to adapt the amount of support as training progresses. The controllers were experimentally tested and evaluation results are presented. The position controller was able to track the desired trajectory, with minimum errors. The path controller avoided leaving the tunnel, but the zero-force controller inside the tunnel allowed subjects to intend the task by themselves. Finally, the guidance-as-needed controller was able to modulate the guidance, allowing more autonomy to initiate the movements as training progressed, while limiting tracking and velocity errors.