Automatic recognition of gait mode from EMG signals of lower limb

This paper describes the application of hidden Markov model (HMM) to recognition of gait mode based on electromyographic (EMG) signals. Four types of time-domain features were extracted from the EMG signals of selected muscles within a time segment. According to the division of the gait phase, the structure of HMM was determined, in which each state is associated with a gait phase. A modified Baum-Welch algorithm was used to estimate the parameter of HMM. And Viterbi algorithm achieved the recognition of gait mode by finding the best HMM and state to assign corresponding phases to the given segments. The locomotion modes could be recognized nearly all correct in evaluation experiments. A satisfactory accuracy in the recognition of gait phase also was obtained.