Gait recognition based on emg information with multiple features

Abstract

In order to evaluate the effects of time domain (TD) and frequency domain (FD) features as well as muscle number on gait classification recognition, eight channels of electromyography (EMG) signals were collected from four thigh and four lower leg muscles, and two TD features and two FD features were extracted in this study. The method of support vector machine (SVM) was presented to investigate the classification property. For the classification stability and accuracy, 3-fold cross validation was verified and selected to classify the lower limb gait. The results show that the FD features can obtain higher accuracy than TD features. In addition, accuracy of gait recognition increased with the augment of muscle number.