Estimation of Elbow Joint Angle Based on Electromyography Using the Sign-Slope Change Feature and Kalman Filtering

Abstract

Electromyography (EMG) has widely used in the field of biomedical engineering as a control signal for prosthetic devices and exoskeleton robotic. High accuracy in the prediction of the limb joint is very important to determine the effectiveness of the system. In this study, we propose a new algorithm to improve the elbow joint angle prediction based on electromyography using Sign Slope Change (SSC) feature extraction and Kalman filter (SSC-KF). The EMG signals acquired from the biceps were extracted using SSC feature to get the estimation of the elbow joint angle. The accuracy of the prediction of the elbow joint angle was improved by using Kalman filter. In this study, the SSC-KF algorithm can predict the elbow joint angle with high accuracy. The Pearson's correlation coefficients (mean±S.D.) were 0.95±0.02, 0.96±0.01, and 0.96±0.015 for the motion period of 12 seconds, 8 seconds, and 6 seconds, respectively. Root Mean Square Errors (mean±S.D.) were 10.37°±1.72°, 9.89°±1.11°, and 9.99°±2.2°, for the motion periods of 12 seconds, 8 seconds, and 6 seconds respectively. This SSC-KF algorithm can predict the elbow joint prediction by using a single lead electrode from biceps muscle.