A HUMAN IDENTIFICATION METHOD BASED ON GAIT CYCLE USING EMG SIGNAL

摘要

The present invention relates to a method to recognize a user based on a walking cycle by using an EMG signal. The method includes: a step (a) of receiving training data for an EMG signal of a walking cycle unit of multiple channels of a user, and generating a classifier recognizing the user with the EMG signal; a step (b) of receiving a test signal for the EMG signal of the walking cycle unit of the channels, and then selecting a classifier in accordance with a recognition rate by calculating out the recognition rate by testing the classifier; and a step (c) of recognizing the user through the selected classifier. The EMG signal of the walking cycle unit is generated by dividing the EMG signal by the walking cycle unit in accordance with a pressure signal of the user. Therefore, the present invention is capable of achieving high security by making forgery difficult while having a constant recognition rate by recognizing the user by using the EMG signal indicating even detailed human movement which is difficult to obtain through a physical sensor.