Preliminary study on analysis of sEMG-based features for force classification

The myoelectric control of upper-limb prosthetic devices is important for the amputee. The pattern recognition to restore control of several degrees of freedom is developed. In this study, the prediction of force level from thumb-index pinch is proposed. The surface electromyography (sEMG) is recorded from three muscle regions (12 channels) of the right forearm of a non-amputee subject. Twelve traditional time-domain features are extracted from collected sEMG signal. The sequential forward floating selection (SFFS) method is investigated to find the optimal set of features. K-nearest neighbor is applied to classify five force levels with three different wrist positions. The results showed that classification accuracy 100% was obtained from 10 features selected by SFFS. The mostly selected group of muscles is from the upper forearm, which also can provide the highest accuracy. This is useful to the prosthetic design for wrist disarticulation amputee and transradial amputee in the future.