Gait Recognition Based on Feature Selection of sEMG Signals Using PCA-RELM Method

To improve the accuracy and real-time performance of gait recognition, this paper studies the gait recognition based on multiple features of surface electromyography (sEMG) signals. Firstly, three types of features, i.e., time domain, frequency domain, and wavelet features, were extracted from the denoised sEMG signals. Then the principal component analysis (PCA) is employed to reduce the dimensionality of the sample features. Finally, three algorithms, i.e., support vector machine (SVM), extreme learning machine (ELM), and regularized extreme learning machine (RELM), are presented in gait recognition, respectively. The results show that the PCA-RELM method can get the higher classification accuracy and recognition efficiency.