EVOLUTIONARY NETWORKS FOR THE IDENTIFICATION OF MYOELECTRIC SIGNALS

摘要

This paper presents initial investigations into an evolutionary neural network suitable for gait analysis of human motion. The aim is to design an intelligent system based on artificial neural networks (ANNs) and evolutionary algorithms, and the possible benefits arising from combinations between them. Evolution was introduced into ANNs using Genetic Algorithms (Gas) evolving connection weights and the architecture of the net. The intelligent black box were able to extract the information stored in each myoelectric signal generated by the muscle and interpret them to give accurate information on the position and movement of the knee (gait). Simulation results are reported for this approach.