sEMG 신호와 주파수 영역 필터를 이용한 보행 모드 분류

초록

This paper proposes a frequency domain filter method to classify the locomotion mode in gait phases using lower limb surface electromyogram(EMG) signals. Force sensors are used to detect before and after of stance and swing phases. For recognizing locomotion mode, EMG signals of all channel are reconstructed and extracted to features in time domain and then they are converted to those in frequency domain by using Fast Fourier Transform (FFT). Locomotion mode is classified by applying frequency domain filters to the feature data. Experimental result shows that average accuracy of classifying locomotion mode at stance and swing phases is 88.1%.