Implementati̇on of system automati̇on based on brai̇n-computer interface by classi̇fyi̇ng of EEG si̇gnals

In this thesis, a Brain-Computer Interface (BCI) is designed, which provides the paralyzed individuals who are not able to use their muscular system and nervous system, can control a system such as wheelchair. In the BCI system, communication of paralyzed individuals with external equipment such as computer, electronics devices and automation systems is provided by interpreting brain activity of the individual. In the BCI technology, brain waves which are called electroencephalography (EEG) are commonly used for measuring the electrical activity of brain, due to the painless measurements and practical considerations. In this study, emotive EEG headset is used for electroencephalography (EEG) recordings of individuals. Output commands of the BCI system which allows controlling of the external devices are generally obtained by applying preprocessing, future extraction and classification procedures to the measured EEG signals. Furthermore, design methods of BCI systems can be different according to EEG signals used in the design. These methods are Visual Evoked Potentials, slow cortical potentials, P300 based BCI systems and μ and β rhythms. In this study, various BCI design methods are examined and then two different BCI systems based on Steady-State Visual Evoked Potentials (SSVEP) and P300 potentials are designed. Speed and direction control of a direct current motor such as clockwise or counter clockwise rotation, acceleration and deceleration is implemented using the designed BCI systems and thus performances of the designed systems are evaluated. In the designed BCI automation systems using the EEG signals measured from the two healthy volunteers, off line control performances are obtained 100% for offline operation and 60%-90% for online operation. Keywords: Brain Computer Interface BCI, EEG signals, Classification, Visual Stimuli, P300, SSVEP.