The Development of an Underwater sEMG Signal Recognition System Based on Conductive Silicon

Abstract

Stroke patients need rehabilitation to recover their abilities of moving, and the underwater rehabilitation can reduce the possibility of secondary injury during the rehabilitation process. Collecting surface electromyography (sEMG) signals underwater can provide better rehabilitation guidance to the medical doctors. However, the current sEMG electrodes cannot be applied to collect sEMG signals underwater. To solve this problem, we propose a soft sEMG electrode based on conductive silicon. The time domain and frequency domain features of sEMG signals are extracted. The sEMG signals are identified by Back Propagation Neural Network (BPNN) under the dry, simulated sweating and water environments respectively. Under the dry environment, there is no significant difference in the recognition accuracy of sEMG signals between the conductive silicon electrode and the Ag/AgCl electrode. Under the water environment, the recognition accuracy of sEMG signals acquired by the conductive silicon electrode is 95.47% by employing time domain features.