An Evolution of Assistive Robot Control to Meet End-User Ability

Abstract

In this work, we present an evolution of system designs and studies that aim to facilitate the operation of high-DoF assistive robotic arms by persons with upper limb paralysis. We highlight the experimental pipeline and note developments in our efforts to design a suitable control map that can convert low variance residual body motions from neuromotor-impaired populations into 6-D velocity control signals for use in teleoperating a 7-DOF robotic arm. Notably, we provide results from variance analyses on raw IMU control signals from both neuromotor-impaired and unimpaired populations, and an analysis of the intrinsic dimensionality of map-building datasets gathered with and without movement guidance. We then present a preliminary 13-session study that vets the control map developed in light of these findings.