Applying Kalman filter in EEG-Based Brain Computer Interface for Motor Imagery classification

Human imagination and intention can be read by using the Electroencephalography(EEG)-Based Brain Computer Interface(BCI) for Motor Imagery. The systems reads the human brain which consists of left or right imaginary movement and decide the actual movement. It is useful especially for the disable people to help their daily life. But the signals include the unwanted signals and other features together with the actual signal. The challenges are to remove the unwanted signals and extract the accurate feature and classify the signal accurately by using the best classification method. This paper applied the Kalman filter to the BCI´s signal processing methods to improve the accuracy and the reliability of the systems. The Common Spatial Pattern (CSP) is used to extract the feature and the Radial Basis Function (RBF) as the classification method. We also compared other classification method which is Fisher Linear Discriminant Analysis (FLDA) to show the RBF gives better result. The simulation result shows improvement by using the proposed method compared to the other method.