Study of Muscular Fatigue Effect on Human-Machine Interface Using Electromyography and Near-Infrared Spectroscopy

Abstract

It is evident that surface electromyography (EMG) based human-machine interface (HMI) is limited by muscle fatigue. This paper investigated the effect of muscular fatigue on HMI performance using hybrid EMG and near-infrared spectroscopy (NIRS). Muscle fatigue inducing experiments were performed with eight subjects via sustained isometric contraction. Four fatigue metrics extracted from EMG and NIRS signals were evaluated during fatigue process. Utilizing the time-varying characteristic of fatigue metrics and their relations, modified features were proposed to dampen the effect of muscle fatigue. The experimental results showed that modified features extracted from combined EMG and NIRS could overcome the impact of muscle fatigue on classification performance to a certain extent, although this slight compensation was still inadequate. It thus suggested that, to minimize the muscle fatigue effect on HMI, high-intensitive sustained muscle contraction should be avoided in HMI usage.