Non-Invasive Hand Free Control of a Robotic Arm

Abstract

Thousands of spinal injuries occur every year, resulting in total or partial quadriplegia. This debilitation robs people of their autonomy, in addition to the large financial cost of care and lost work. Some devices to return some autonomy exist, however, most of them require some movement or involve pre-programmed actions for a set environment. This experimental study seeks to design and compare novel methods of controlling a robotic arm pointer to a baseline, hand-controlled, mode. The modalities compared include a heads-position using a motion sensor to directly map the tip of the robotic arm to the position of the head, head velocity to create a vector to control the direction and speed based on the position of the head, and voice control which causes the tip to move in a singular direction or stop based on specific vocal input. Head position was found to be similar to hand control, performing significantly better than head velocity and voice control when observing movement time to target and throughput. Path length saw no significant differences between baseline and the three experimental modalities, and the NASA TLX showed a noteworthy dislike for head velocity mode. While this study lacks any form of gripping mechanism, it lays the groundwork for head position mode to be a novel method of control, for individuals with partially or totally limited body movement.