Development of a Hybrid Surface EMG and MMG Acquisition System for Human Hand Motion Analysis

Surface electromyography (EMG) is widely investigated in human-machine interface (HMI) by decoding movement intention to intuitively control intelligent devices. Mechanomyogram (MMG) is the mechanical vibration signal produced by contracting muscles, and is also useful tool for intuitive HMI. Combining EMG and MMG together would provide more comprehensive information of muscle activities. This paper presents a hybrid EMG and MMG acquisition system to capture these two signals simultaneously. Experiments were carried out to study the performance of the proposed hybrid sensor for human hand motion analysis, and the results indicated that EMG and MMG reflected muscle contractions from different aspect for their different time-frequency responses. Furthermore, the pattern recognition experiment results showed that the classification accuracy using combined EMG-MMG outperformed EMG-only by 20.9% and 14.6%, with mean absolute value (MAV) and power spectral density (PSD) features respectively. The findings of this study support and guide the fusion of EMG and MMG for improved and robust HMI.