Change search
ReferencesLink to record
Permanent link

Direct link
A Fuzzy Logic Classifier Design for Enhancing BCI Performance
University of Ulster. (School of Computing and Intelligent Systems)ORCID iD: 0000-0001-6553-823X
University of Ulster. (School of Computing and Intelligent Systems)
University of Ulster. (School of Computing and Intelligent Systems)
2006 (English)Conference paper, Poster (Refereed)
Abstract [en]

This work is aimed at enhancing inter-session performance of Brain-Computer Interface (BCI) classification. The effective handling of uncertainties associated with changing brain dynamics is considered to be a key issue. Since fuzzy logic (FL) has been recognized as a functional and well-suited approach to capturing the effects of uncertainty, the research has been concentrated on the development of an FL classifier for a BCI system. The emphasis is placed on type-2 (T2) FL methodology that has recently emerged as an expanded version of classical type-1 (T1) FL. In this work a case study was conducted using ECoG recordings made available as part of BCI competition III. Due to high dimensionality of the signal, two-stage feature selection was devised. The overall performance of the developed BCI was assessed in off-line simulations based on the classification accuracy (CA). Comparative analysis of the designed T2FL and T1FL systems with LDA as BCI classifiers suggests that T2FL has superior capability in effective dealing with inter-session variability of the ECoG dynamics in the given subject.

Place, publisher, year, edition, pages
Keyword [en]
brain-computer interface, electroencephalogram, fuzzy systems, classifier
National Category
Information Systems
URN: urn:nbn:se:kth:diva-90102OAI: diva2:504140
3rd International Brain-Computer Interface; Workshop and Training Course 2006. Graz, Austria. September 2 1-24 2006
QC 20120529Available from: 2012-02-19 Created: 2012-02-19 Last updated: 2012-05-29Bibliographically approved

Open Access in DiVA

No full text

Search in DiVA

By author/editor
Herman, Pawel Andrzej
Information Systems

Search outside of DiVA

GoogleGoogle Scholar
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

Total: 47 hits
ReferencesLink to record
Permanent link

Direct link