Interactive gesture controller for a motorised wheelchair

Sciberras, Jade (2015) Interactive gesture controller for a motorised wheelchair. Honours thesis, Murdoch University.

Abstract

This paper explores in great detail the design and testing of a gesture controller for a motorised wheelchair. For some, motorised wheelchairs are part of their everyday life. Those individuals who depend on their motorised wheelchair do so for a vast range of reasons; therefore, it is reasonable to assume that modifying and improving upon the standard joystick controller for a motorised wheelchair can benefit a person’s way of life significantly.

The design of the gesture controller is heavily based around the user’s needs so as to benefit them and compliment their strengths to give them more control. For individuals with limited movement and dexterity, the user interface, system responsiveness, ergonomics and safety were considered when engineering a system that is intended for people to use.

A device capable of recognising a hand gesture was carefully chosen. The technology that is readily available for this application is relatively new and not extensively documented. The LEAP motion sensor was chosen as the hand gesture recognition device to be the controller for a wheelchair. This device has hand recognition software but the device’s software lacks the predictability and accuracy required for a motorised wheelchair controller. Through testing, the controller accuracy improved. Although this controller is adequate for a laboratory environment, further testing and development will be required for this alternative wheelchair controller to evolve into a commercial product.

The gesture triggered controller was designed around the capabilities of the developer’s hand; but the method outlined in this paper is transferable to any individual hand size and more importantly the limitations of their hand gestures. The outcome of this thesis is a customised, non-invasive hand gesture controller for a motorised wheelchair that is able to be fully tailored to a person’s capability without losing it responsiveness or accuracy.

Item Type:	Thesis (Honours)
Murdoch Affiliation(s):	School of Engineering and Information Technology
Supervisor(s):	Parlevliet, David
URI:	http://researchrepository.murdoch.edu.au/id/eprint/29873

Item Control Page

Downloads