Classification of Myoelectric Signals Using Multilayer Perceptron Neural Network with Back Propagation Algorithm in a Wireless Surface Myoelectric Prosthesis

Abstract

The paper focuses on a wireless myoelectric prosthesis of the upper-limb that uses a Multilayer Perceptron (MLP) neural network with back propagation algorithm in classifying electromyography (EMG) signals. MLP Neural network is composed of processing units that have the capability of sending signals to each other and perform a desired function. The algorithm is widely used in pattern recognition. The network is used to train EMG signals and use it in performing the necessary hand positions of the prosthesis. Through programming a Field Programmable Gate Array (FPGA) using Verilog and transmission of data with Zigbee, the EMG signals are acquired, classified, and simulated wirelessly. The signals are classified and trained to produce the necessary hand movements. The corresponding hand movements of Open, Pick, Hold and Grip are simulated through the Zigbee controller. Z-test is used to analyze the data that were produced and acquired from using the neural network.