Hägusate klassifikatsioonialgoritmide kasutamine aju-arvuti liidestes
Abstract
Selles lõputöös uuritakse hägusate klassifikatsioonialgoritmide kasutamist elektroentsefalograafial (electroencephalography, EEG) põhinevates aju-arvuti liidestes (brain-computer interfaces, BCI). Uuritakse olemasolevat kirjandust BCI-des kasutatavate klassifikatsioonialgoritmide, hägusate algoritmide olemuse ja nende kasutamise kohta BCI-des. Hägusate algoritmide potentsiaalsete eeliste demonstreerimiseks realiseeritakse lihtne aju-arvuti liides, mis võimaldab kasutajal liigutada kursorit arvuti ekraanil. Testid selle rakendusega näitavad, et hägusad algoritmid sellist tüüpi rakendustes ei oma eelist traditsiooniliste algoritmide üle.
In this thesis, the usage of fuzzy classification algorithms in brain-computer interfaces (BCI) based on electroencephalography (EEG) is researched. We review the existing literature on BCI, the traditional crisp algorithms often used in BCI for classification, fuzzy classification algorithms and their application in BCI. A simple BCI system is implemented that allows the user to move a cursor on the computer screen. Tests conducted with this application show that fuzzy classification algorithms do not have advantage over crisp classification algorithms in this kind of BCI systems.
In this thesis, the usage of fuzzy classification algorithms in brain-computer interfaces (BCI) based on electroencephalography (EEG) is researched. We review the existing literature on BCI, the traditional crisp algorithms often used in BCI for classification, fuzzy classification algorithms and their application in BCI. A simple BCI system is implemented that allows the user to move a cursor on the computer screen. Tests conducted with this application show that fuzzy classification algorithms do not have advantage over crisp classification algorithms in this kind of BCI systems.