Abstract
This study developed a motion discrimination method considering variation of EMG signals associated with lapse of time. In previous study, we proposed real-time discrimination method based on EMG signals of forearm. Our method uses hypersphere model as a discriminator. In motion discrimination using EMG signals, there is a problem to maintain high discrimination accuracy over time. It's known that EMG signals changes along with lapse of time. This study analyzed the effect of changes in EMG signals to our method. From result of the analysis, we propose adding relearning system of decision criteria to the discrimination system. A motion discrimination method that contains the relearning system was created and the effectiveness of adding relearning system was experimentally verified. The motion discrimination system was able to discriminate 3 hand motions with discrimination accuracy above 90 % and discrimination processing time below 300 ms even after the time has elapsed.
