Abstract
Many forearm amputees use various prosthetic hands depending on their purposes. Myoelectric prosthetic hands, which use electromyogram (EMG) signals in persisting muscles as operation signals, are attracting attention because they look similar to cosmetic prosthetic hands and can be operated by a sense of nature. The objective of this study is to develop a method of discriminating real-time motion from EMG signals. We previously proposed a real-time motion discrimination method using the hyper-sphere models. This method could discriminate between five motions (hand opening, hand closing, hand chucking, wrist extension, and wrist flexion) at a rate of above 90 percent. The discrimination processing time made was less than 300 ms by using the hyper-sphere models. Moreover, this method prevents elbow motions from interfering with hand motion discrimination. However, forearm amputees cannot go about their daily life comfortably by using only five motions. Furthermore, we think that wrist pronation and wrist supination are important in daily life. Here, we propose a real-time motion discrimination method using the hyper-sphere models for seven motions: hand opening, hand closing, hand chucking, wrist extension, wrist flexion, wrist pronation, and wrist supination. Furthermore, we also created new hyper-sphere models to maintain real-time motion discrimination. We experimentally verified that the discrimination accuracies of this method were 90 percent and the discrimination processing time was less than 300 ms. Moreover, elbow motions did not interfere with the hand motion discrimination.
