Fusion of myoelectric and near-infrared signals for prostheses control

Abstract

Assistive tools for mobility and manipulation require sophisticated control interfaces, tailored to the capacity of the user while compensating limitations, such as muscle fatigue. The key to control-transparency for the patient, safety and security during operation, and real-time performance of the assistive device are astute sensors. In this contribution we bring forward a novel sensor, in which we combine a myoelectric and a near-infrared sensor, to detect muscle exertions for the real-time control of assistive devices. Since our research of over more than a decade now focuses on the electronic control of prostheses for the upper-extremity, as used by patients with either above- or below elbow amputations, we provide solid examples from this domain. As the presented technology can clearly be used to control also many other assistive devices, we predominantly indicate it for individuals who cannot directly use body-powered devices for their daily activities. For the successful application of our sensors, a patient is only required to willingly isolate the contraction of extant muscles, for which a minimal-threshold myoelectric or reflected near-infrared signal can be measured cutaneously. Experiments with the miniature sensors show, that the combined sensor works very well and the spatial resolution surpasses that of existing devices. Furthermore, the fused signal can be readily used as input to a classifier and allows to adjust for the effects of muscle fatigue.