Desenvolvimento de uma interface cérebro-computador não invasiva baseada em potenciais evocados visuais de regime permanente aplicada à comunicação alternativa e robô de telepresença

Abstract

There are people who have their communication skills impaired due to infirmities, accidents or injuries. New technologies are emerging for those people, in order to provide them an alternative communication through brain signals. These systems are known as Brain Computer Interfaces (BCIs). This thesis describes the development of a SSVEP (Steady State Visual Evoked Potential) based BCI applied to alternative communication and that can also be used in a telepresence robot. The interface allows the choice of four commands using visual stimuli, which were developed using the OpenGL graphical library and that runs in distinct frequencies (5.6Hz, 6.4Hz, 6.9Hz and 8.0Hz). All volunteers who participated the tests using the online system could complete the tasks. The average rate of success was 88.3% ± 5.4% and the classification time was 5.6s ± 0.5s, with mean ITR of 14.2 bits/min ± 3.5 bits/min, without the need of training and using a single channel. The results show the possibility of building a BCI that can be used in future projects of assistive technologies in the Intelligent Automation Laboratory at University of Espirito Santo.