Simplification of EMG Feature Extraction System Using Digital Signal Processor

摘要

This paper reports a forearm motion discrimination system for control of myoelectric hand prostheses. Simplification of EMG feature extraction system is important to design small-sized prosthesis control system. The EMG feature extraction system using DSP was evaluated with the DSP starter kit (6711DSK). Feature vectors were obtained using FFT function. Owing to the EMG frequency shift, values of each element of the vectors will exchange. It causes incorrect discriminating. We perform the reduction of the incorrect discriminating using 12-period moving average.