Using Artificial Neural Network to Model EMG Signals from the Prime Movers of the Shoulder for Rehabilitation Robotic Systems

Authors

  • D. Matheson Rittenhouse School of Engineering, University of Guelph Guelph, Ontario, CANADA N1G 2W1
  • Hussein A. Abdullah School of Engineering, University of Guelph Guelph, Ontario, CANADA N1G 2W1
  • R. John Runciman School of Engineering, University of Guelph Guelph, Ontario, CANADA N1G 2W1
  • Otman Basir School of Engineering, University of Guelph Guelph, Ontario, CANADA N1G 2W1

Abstract

Artificial neural networks have demonstrated some ability to model the electromyogram (EMG) signals from prime movers of the joint under investigation. This paper demonstrates the ability of a fully connected feed forward neural network (FF NN) to predict EMG signals from eight muscles of the shoulder. Robots used for physical rehabilitation can incorporate the information from EMG as an input to an intelligent decision making algorithm used to adjust the level of difficulty according to patient performance.

Downloads

Published

2002-12-31

How to Cite

[1]
D. M. Rittenhouse, H. A. Abdullah, R. J. Runciman, and O. Basir, “Using Artificial Neural Network to Model EMG Signals from the Prime Movers of the Shoulder for Rehabilitation Robotic Systems”, CMBES Proc., vol. 27, no. 1, Dec. 2002.

Issue

Vol. 27 (2002)

Section

Academic