Linear and Non-Linear Classification of EMG Signals for Probable Applications in Designing Control System for Assistive Aids

Uvanesh, K (2015) Linear and Non-Linear Classification of EMG Signals for Probable Applications in Designing Control System for Assistive Aids. MTech thesis.

[img]
Preview
PDF
3854Kb

Abstract

EMG signal was acquired by placing electrodes on the surface of forearm muscle. The acquisition is made possible using a bio-potential amplifier (Gain ˜ 2500 with a cut off frequency of 1500Hz). The acquired EMG signal was processed further, so that the EMG signal can be classified into their corresponding category.[1] By using the raw EMG signal, the envelope of the signal were detected, then original EMG signal were extracted, later the extracted EMG signal was Wavelet processed. For preforming the classification, the features were extracted. By using the extracted features, Offline and Online classifications were performed. The results showed an accuracy of >95% (overall). For improving the performance of the classification, Boolean change state logic and Hall Effect sensor were used to design the control system.

Item Type:Thesis (MTech)
Uncontrolled Keywords:EMG signal, Envelope creation, Extracted signal, Wavelet Processed extracted signal, Offline classification, Online Classification
Subjects:Engineering and Technology > Biomedical Engineering
Divisions: Engineering and Technology > Department of Biotechnology and Medical Engineering
ID Code:7985
Deposited By:Mr. Sanat Kumar Behera
Deposited On:23 Jun 2016 18:08
Last Modified:23 Jun 2016 18:08
Supervisor(s):Pal, K

Repository Staff Only: item control page