Title: Effect of muscle length changes on classification of EMG for prosthesis control

Authors: Nadipi Reddy Prabhav; Suresh R. Devasahayam

Addresses: Department of Bioengineering, Christian Medical College, Bagayam, Vellore, Tamil Nadu 632002, India; School of Biosciences and Technology, Vellore Institute of Technology, Vellore, Tamil Nadu 632014, India ' Department of Bioengineering, Christian Medical College, Bagayam, Vellore, Tamil Nadu 632002, India

Abstract: Significant advances in the use of pattern classification-based Electromyography (EMG) control for prosthesis have not resulted in clinical solutions. This paper explores whether muscle length-dependent changes in EMG characteristics play a significant part in this. We show, theoretically and experimentally, that the EMG frequency spectrum is affected by fibre length change. Further, we use this EMG in conjunction with a clustering algorithm to show that this length dependence has a profound effect on the classification of EMG. The clustering is done by extracting the wavelet feature from the EMG and using the k-means clustering algorithm. Various ways of compensating for this muscle length dependence of the EMG are suggested.

Keywords: prosthesis control; electromyography; EMG electrode configuration; EMG classification; EMG clustering; muscle length dependence; muscle length changes; pattern classification; k-means clustering; wavelets.

DOI: 10.1504/IJBET.2013.057929

International Journal of Biomedical Engineering and Technology, 2013 Vol.13 No.2, pp.117 - 132

Received: 25 Feb 2013
Accepted: 07 Oct 2013
Published online: 27 Sep 2014 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article