An enhanced human-computer interface based on simultaneous sEMG and NIRS for prostheses control

Surface electromyography (sEMG) is extensively investigated in human-computer interface (HCI) for prostheses control to improve the the life quality of the amputees. Near-infrared spectroscopy (NIRS) reflecting muscle activity in hemo-dynamics and metabolism, is less explored in HCI applications. Reasonably combining the advantages of both sEMG and NIRS would provide a novel approach to enhance the sEMG based HCI performance. This paper presents an enhanced HCI via decoding sEMG and NIRS simultaneously. Pattern recognition experiment of thirteen motion classes is performed with seven feature sets extracted and compared. The results show that the classification accuracy is significantly (P<;0.001) improved by using combined sEMG and NIRS features comparing to sEMG or NIRS features individually. Providing more information about the muscle activities, fusion of sEMG and NIRS is available for an enhanced HCI for prostheses control.