Research on Gait Recognition Based on Lower Limb EMG Signal
Junyao Wang, Yuehong Dai, Kang Tong, Xiaxi Si
Abstract
To verify whether the EMG signals can recognize different gait under different road conditions, this paper selects 5 kinds of gait (walking on flat ground, uphill, downhill, upstep and downstep), collects the EMG signals of 4 muscles (Biceps femoris, lateral femoral, tibialis anterior, and gastrocnemius) of lower limbs during the movement, and analysis the differences of the peak values of EMG signals under different gait conditions. To compare the recognition effect, the thigh and calf EMG signals were used as input to recognize 5 kinds of gait. The results show that lower limb EMG signals can recognize different gait. The recognition rate for uphill and downhill gait by thigh EMG signal is low (75.37% and 72.20%), and that in calf EMG signal for uphill gait is low (73.94%), but when the 4 muscles of thigh and calf are input together, the recognition rate of the five gait is improved, and the average recognition rate is 95.16%.