Control Method Based on EMG for Power Assist Glove Using Self-Organizing Maps

Abstract

In this study, electromyography is applied to a subject’s forearm to discriminate between intended finger and wrist motions and the resulting signals are used to control a power assist glove. This glove is designed to assist with grasping. The fingers and wrist are moved when the corresponding muscles of the forearm contract simultaneously. Therefore, the intention to move either a finger or a wrist must be discriminated to enable assisting a finger movement by this glove. The introduction of Self-Organizing Maps (SOM) improves the ability to discriminate between these motions. In this study, the effectiveness of a discrimination method using SOM is verified experimentally.