Real-time identification of electromyographic signals from hand movement

In this work, we have demonstrated a novel on-line technology for real-time acquisition and identification of electromyographic (EMG) signals from hand movement. EMG signal were measured using standard surface electrodes from forearm muscles at three major points including Wrist extensor, Flexor Carpi Radialis and Wrist Flexor groups, respectively. The EMG acquisition system consists of an instrumentation amplifier, filter circuit, isolator, an amplifier with gain adjustment and a commercial embedded system called FiO board. The commercial FiO embedded system is interfaced with the computer and EMG is represented, analyzed and stored in real-time on computer by Simulink program. EMG signals are identified by RMS and SD feature extraction methods and k-mean clustering algorithm. The result revealed that both RMS and SD can be used with k-mean clustering algorithm to obtain the distinct Euclidean distance characteristic of EMG signal for each movement. The minimum Euclidean distance with RMS and SD for each hand movement uniquely occurs at a distinct Euclidean distances between real EMG data and extracted features.