2채널 바이폴라 뇌파센서 기반의 운동심상을 활용한 BCI 시스템 구현

Abstract

EEG is a measurement technology for electrical signals generated by reaction activity of the brain and it had been used in many researches for study of the brain which is the most complex part in the human body. EEG signal can represent human’s thought and will, so the research in Brain-Computer Interface(BCI) which is the control interface of the new paradigm using EEG has gained attention recently. In this study, BCI system is constructed using 2 channel bipolar EEG sensor device developed by ourself. The EEG sensor device consists of analog circuit and digital circuit. Analog circuit is designed for noise removal and signal amplification of raw EEG signal which are measured through electrode, while digital circuit is designed for conversion of the analog data from analog circuit into digital data and construction of data packet for transmission purpose. Data packet is transmitted by using Bluetooth wireless communication to the windows-based application for classification of user’s intention. In this study, motor imagery(MI) which is one of the method among several analysis methods in BCI system is implemented for classification. In windows-based application, Fast Fourier Transform(FFT) is implemented to obtain a Power Spectral Density(PSD) value of EEG data. Support Vector Machine(SVM) performs MI classification by using PSD as the feature vector. EEG data group for feature extraction is collected by using training program which is designed for selection of clear MI sample data. Real time feedback application which consists of virtual control system and driving control system is developed to analyse user intention. The proposed BCI system not only provides an alternative control interface for disable people and patient, it also offers end user a convenient and fast user interface.