Research of Gait Recognition Based on Human Electrostatic Signal

The electrostatic induction signal could be sensed when charged human body was walking, and discrepancy of walking habits such as gait frequency, stride of different individuals will be reflected in the signal. This paper proposed a method to recognize human gait by measuring electrostatic signal. A modified dynamic time warping algorithm was utilized to recognize human gait after de-noising the electrostatic signal which was acquired by a non-contact induction electrode. In order to verify the performance of this method, we collected 25 gait signal segments from 5 subjects, 5 signal segments for each subject. The result demonstrates that our method has a correct rate over 85%, and greatly shortens the operative time compared with the traditional DTW algorithm.