Development of an EMG-based Human-Machine Interface on Open-source Linux Platform for Evaluating the Motor Skill Acquisition Process

摘要

Human-machine interface (HMI) not only assists disabled people to perform tasks and activities, but also helps in acquisition of motor skills for rehabilitation. Therefore, it is critically important to qualitatively assess the motor skill acquisition process during the operation of HMI. Some of the recent studies observed and analyzed that changes in bio-signals were closely related to the improvement of motor skill. This paper presents a HMI for balancing the control of virtual inverted pendulum using electromyography (EMG) signals. The proposed HMI was developed using an open-source control software RTAI (Real Time Application Interface) on Linux platform. EMG signals from the forearm were recorded using two EMG electrodes placed around the extensor carpi radialis longus and flexor carpi ulnaris muscles. Moreover, the motor skills were evaluated by classifying the EMG signal patterns (i.e., muscle activation patterns), and measuring the motion of inverted pendulum including angular position, speed, and acceleration. An experiment was conducted using the proposed HMI with 8 subjects in laboratory. The results showed that the proposed HMI is useful to examine the motor skill acquisition process, as the EMG signal patterns closely resembled to that of human motor process.