Detection of Human Action Intention by Electromyography (EMG)

Abstract

Due to the growing proportion of elderly people in modern society, there is significant need for an increased understanding of the mechanisms of skeletal muscle structures and functions. This will allow robots to be developed to provide support for the elderly. Wearable robots have been developed for applications related to assistance and rehabilitation. The control strategy of a wearable robot is a subtle balance between the generation of muscle force from a human and the generation of force from a robot assistant. The intention of standing should be detected instantaneously, as the time window for the effect of assistance is much shorter. According to the literature, electronic signals generated from muscles are present before the force itself. Therefore, surface electromyography (sEMG) is a suitable approach for detecting the action of human movement. An approach for detecting human action intention based on sEMG is presented in this paper. By using the Kalman Filter, the response time can be shrunk within 20ms, which will provide an instant signal to the robot control system and improve the robot assistant's performance.