A discrimination system using neural network for EMG-controlled prostheses

The electromyographic (EMG) signal from active muscle is observed on the surface of the living body, and considered for controlling an externally powered upper extremity prosthesis. However, the EMG signal depends on physical condition, the state of mind and so on. So it is difficult that the original EMG signal could be used as a command for controlling an externally powered upper extremity prosthesis directly. In this paper, focusing on excellent functions of neural network of learning and processing, a discrimination system using a neural network for generating commands to control EMG-controlled externally powered upper extremity prosthesis is proposed. The neural network is used in this system to learn the relation between the power spectrum of the EMG signal analyzed by fast Fourier transform method and the performance desired by the handicapped. The neural network has three layers, that is, the input layer, the middle layer and the output layer. It was cleared that the discrimination system with the neural network could discriminate 7 performance from the EMG signals with the probability of 0.81