Design Considerations for Wireless Acquisition of Multichannel sEMG Signals in Prosthetic Hand Control

Abstract

Wearable technology for assistive medical applications and physical activity recognition has emerged as a fast growing research field in recent years. However, the design of such systems still poses challenges, including restricted physical size, limited computational resources, and the availability of constrained energy sources. In this paper, we present a practical design space exploration of a body-worn system for prosthetic hand control based on surface electromyography (sEMG) signals. The sEMG method is a well-established sensing technology that provides the detection of electrical activity produced by the physiological contractions of muscles. The presented wearable system is designed to acquire sEMG signals for successive recognition of the performed motion and the control of the prosthetic hand, allowing to regain a considerable amount of life quality for a broad patient community. The main guiding requirements for the presented wearable system include real-time data acquisition from up to 32 sEMG channels, reliable wireless data streaming, long mobile autonomy (several days lifetime), non-intrusive mounting, and compact size. We present the system's hardware and software architecture focusing on the comparison of various communication design options such as recent Bluetooth low energy and low-power WiFi technologies.