Study And Implementation Of Brain Computer Interface System Based On EEG

The Brain Computer Interface(BCI) research has drawn attention of scientists in brain science, rehabilitation engineering, biomedical engineering and intelligent information processing. The BCI system aims at creating new direct information interaction and communication channels between brain and computer,rather than depending on brain’s normal output channels of peripheral nerves and muscles.In this paper, we mainly focus on Motor Imagery(MI) Based BCI systems based on electroencephalogram(EEG) signals from which several MI tasks can be recognized. We mainly research on several aspects including experiment design, neurophysiology mechanism, feature extraction algorithms and online BCI systems.This Paper proposed a Pre-Processing method which combines the Independent component analysis(ICA) and Common spatial pattern(CSP).In EEG recordings, various noises and interference would be added to the EEG signals, especially ocular artifacts which usually has high amplitude. ICA was used to separate EEG signals to obtain the independent components, then a parts of independent components was selected based on the prior knowledge toreconstruct motor-image related EEGfeature signals.Common spatial pattern(CSP) algorithm was used to extra the spatial featurebetween two types of motor imaged tasks, which build a spatial filter, tomaximum the differencesof variances between the two types of projected feature signals, and then fisher linear classifiers and support vector machine were used to recognize the different labels.Finally, we designed a Pioneer 3 Robot control system based on BCI, including its hardware and software parts. Then a large number of experiments were implemented to investigate the performance of the system. Experimental results provedthe validity of the feature extract method and optimize method, which paved the wayto the on-line BCI system’s application in the future.