ARTIFICIAL LIMB MOVEMENT SYSTEM USING K-STAR ALGORITHM - A DISCRETE WAVELET TRANSFORMATION APPROACH

Abstract

Machine learning is one of the promising areas which contributes to human rehabilitation and this statement can be substantiated by the number of researches conducted in this field. In this study, we attempt to direct an artificial limb system with the help of Electroencephalogram (EEG) signals. EEG signals are created as a result of brain activities when humans intend to perform any action. Hence, capturing this signal and using them to control the artificial limb will be as close to how a human will control their normal hand. Four separate “classes of EEG signals were recorded from” 27 healthy subjects while they were instructed to perform various hand movements such as Finger open (Fopen), Finger close (Fclose), Wrist counterclockwise (WCCW) and Wrist clockwise (WCW). The recorded EEG signals were further classified with classification algorithm to identify the desired movement. Feature extraction, feature selection and feature classification are the three important phases of machine learning which needs to be focused on. The aim of this study is to mine Discrete Wavelet features from EEG signals, classify them with K-Star algorithm, and propose the best features that can be used to regulate the artificial limb.