Trajectory Control for a Myoelectric Prosthetic Wrist

Abstract

We present a novel method for controlling a myoelectric prosthetic wrist. Five multiple degree-of-freedom (DOF) wrist trajectories are obtained from healthy participants that performed tasks that span the range of Activities of Daily Living (ADL) using dimensionality reduction and unsupervised machine learning techniques. The efficacy of these motions is tested as part of a pilot study where a participant used a simulated wrist device controlled using two-site surface electromyography (sEMG); two trajectories were tested in an immersive virtual reality. Novel wrist control has been demonstrated to be more intuitive to use and appears more natural while limiting the amount of body compensation.