Mechanical control of a wheelchair by means of EEG signals

Date
2018-03
Journal Title
Journal ISSN
Volume Title
Publisher
Stellenbosch : Stellenbosch University
Abstract
ENGLISH ABSTRACT: The aim of this project was to incorporate a low cost electric wheelchair by modifying an existing self-propelled one with a neuro headset. The headset must control the wheelchair by using EEG data and motor imagery. Binary and discriminant analysis classifiers were implemented with accuracies ranging between 44% and 68.75%. It is concluded that dynamic classification has a very low accuracy compared to using pre acquired data. Also motor imagery is not very well suited when used with the Epoc+ neuro headset. The wheelchair was never constructed, however it was concluded that the listed components are sufficient to create a cheaper alternative for an electric wheelchair.
AFRIKAANSE OPSOMMING: Die doel van die projek was om `n lae koste elektriese rolstoel te inkorporeer deur `n bestaande stoot-rolstoel te modifiseer. Hierdie rolstoel moet deur `n neuro-kopstuk beheer word deur middle van EEG seine en motor beelde. Binêre en diskriminante analise klassifikasies was gebruik met akkuraathede tussen 44% en 68.75%. Hieruit word afgelei dat dinamiese klassifikasie minder akkuraat is as wanneer vooraf bepaalde waardes gebruik word. Ook is motor beelde nie `n baie goeie keuse wanneer die Epoc+ neuro-kopstuk gebruik word nie. Die rolstoel was nooit gebou nie, maar daar is afgelei dat die componente gelys in die tesis voldoende is om `n goedkoop alternatief vir `n elektriese rolstoel te bou.
Description
Thesis (MEng)--Stellenbosch University, 2018.
Keywords
EEG, Wheelchairs, UCTD, Mechanical movements
Citation