Biometric Personal Identification Based on Gait Analysis Using Surface EMG signals

Abstract

Personal identification based on EMG signals have a robust security, this is due to the fact that people have unique characteristics seen in their walking habits. Such habits include walking stride, speed, duration of burst and magnitude of the EMG signal from their muscles. This paper presents a novel personal identification method by analyzing gait habit using EMG signals from the low-limb muscles of an individual. We designed the personalized matrix of features and muscles for individuals. For this system, five-dimensional training and test data are extracted from the values of root mean square (RMS), mean absolute value (MAV), dominant frequency (DF), integrated EMG signals (iEMG), and onset time of muscle firing(OS). Experimental results show that the average identification accuracy value of the proposed method, in terms of the personal identification with personalized matrix of features and selected muscles, is 93%. This was determined using the linear discriminant analysis (LDA) classifier.