Regression Based Real Time Hand Gesture Recognition and Control for Electric Powered Wheelchair

Iqbal, H; Zheng, J; Chai, R; Chandrasekaran, S

Regression Based Real Time Hand Gesture Recognition and Control for Electric Powered Wheelchair

Iqbal, H Zheng, J Chai, R

Chandrasekaran, S

Permalink

Publication Type:: Conference Proceeding
Citation:: Australasian Conference on Robotics and Automation, ACRA, 2022, 2022-December
Issue Date:: 2022-01-01

Closed Access

	Filename	Description	Size
	pap119s2.pdf	Published version	11.22 MB	Adobe PDF	View/Open

Copyright Clearance Process

Recently Added
In Progress
Closed Access

This item is closed access and not available.

Full metadata record

Field	Value	Language
dc.contributor.author	Iqbal, H
dc.contributor.author	Zheng, J
dc.contributor.author	Chai, R https://orcid.org/0000-0002-1922-7024
dc.contributor.author	Chandrasekaran, S
dc.date.accessioned	2023-05-12T21:30:26Z
dc.date.available	2023-05-12T21:30:26Z
dc.date.issued	2022-01-01
dc.identifier.citation	Australasian Conference on Robotics and Automation, ACRA, 2022, 2022-December
dc.identifier.isbn	9781713866480
dc.identifier.issn	1448-2053
dc.identifier.uri	http://hdl.handle.net/10453/170324
dc.description.abstract	Steering an electric-powered wheelchair is an onerous task for a paralyzed person. Hence, there is a need for either designing a new one or modifying the existing electric-powered wheelchair that is intelligent enough and provides easy daily use for a person who is not capable of handling the manual steering process. Our proposed system is designed to receive, process and classify the surface electromyography (sEMG) signals and gesture recognition techniques before controlling the wheelchair. This paper is based on an analysis of sEMG signals and gesture recognition techniques of a user's dominant limb, and its deployment through Artificial Intelligence based machine learning algorithms. In myoelectric control, classification has been showing promising results with high accuracy but is well known for non-intuitive control. The regression model, on the other hand, allows human-like natural movements, producing proportional and simultaneous control. We are using hand gesture control of an unidirectional wheelchair using sEMG as wearable sensors. Five basic gestures are recognized and classified using feature extraction and a machine learning algorithm. These gestures are mapped to the unidirectional motion commands to steer the wheelchair. The classified algorithm and realtime navigation of the smart wheelchair using the proposed algorithm have been tested by 6 healthy subjects. The results demonstrate performance improvement and gesture recognition accuracy of 95.50% and reduced training time (< 2 mins), compared to state-of-art regression models. In addition, this algorithm has been applied to proportional and simultaneous myoelectric control in real-time.
dc.language	en
dc.relation.ispartof	Australasian Conference on Robotics and Automation, ACRA
dc.rights	info:eu-repo/semantics/closedAccess
dc.title	Regression Based Real Time Hand Gesture Recognition and Control for Electric Powered Wheelchair
dc.type	Conference Proceeding
utslib.citation.volume	2022-December
pubs.organisational-group	/University of Technology Sydney
pubs.organisational-group	/University of Technology Sydney/Faculty of Engineering and Information Technology
pubs.organisational-group	/University of Technology Sydney/Faculty of Engineering and Information Technology/School of Electrical and Data Engineering
pubs.organisational-group	/University of Technology Sydney/Centre for Health Technologies (CHT)
utslib.copyright.status	closed_access	*
dc.date.updated	2023-05-12T21:30:19Z
pubs.publication-status	Published
pubs.volume	2022-December

Abstract:

Steering an electric-powered wheelchair is an onerous task for a paralyzed person. Hence, there is a need for either designing a new one or modifying the existing electric-powered wheelchair that is intelligent enough and provides easy daily use for a person who is not capable of handling the manual steering process. Our proposed system is designed to receive, process and classify the surface electromyography (sEMG) signals and gesture recognition techniques before controlling the wheelchair. This paper is based on an analysis of sEMG signals and gesture recognition techniques of a user's dominant limb, and its deployment through Artificial Intelligence based machine learning algorithms. In myoelectric control, classification has been showing promising results with high accuracy but is well known for non-intuitive control. The regression model, on the other hand, allows human-like natural movements, producing proportional and simultaneous control. We are using hand gesture control of an unidirectional wheelchair using sEMG as wearable sensors. Five basic gestures are recognized and classified using feature extraction and a machine learning algorithm. These gestures are mapped to the unidirectional motion commands to steer the wheelchair. The classified algorithm and realtime navigation of the smart wheelchair using the proposed algorithm have been tested by 6 healthy subjects. The results demonstrate performance improvement and gesture recognition accuracy of 95.50% and reduced training time (< 2 mins), compared to state-of-art regression models. In addition, this algorithm has been applied to proportional and simultaneous myoelectric control in real-time.

Please use this identifier to cite or link to this item:

http://hdl.handle.net/10453/170324