Fusion Eye Multitask Mode Of Electric Information Research And Implementation Of Brain-computer Interface System

Hybrid brain computer interface(HBCI)is constructed by two or more BCI systems with different brain activity pattern or different input signal sources,such as another kind of EEG signal mode or heart rate,eye movement,blood flow changes and other physiological signals.HBCI has many advantages,such as many kinds of tasks,high recognition accuracy,flexible control mode and so on,which has become the new direction to improve the performance of BCI system.In this paper,an online control system of hybrid brain computer interface which comprise of motor imagery EEG and eye movement systems is studied.Our works included data acquisition,preprocessing,feature extraction and pattern recognition methods of EEG and EOG signals.Finally the real time control flight experiment of the four axis aircraft was completed.The main research work of this paper is described as follows:(1)Based on the analysis of feature extraction methods of time domain analysis,time frequency analysis and common spatial pattern,OVO-CSP and OVR-CSP algorithms are designed to extract the features of multi class EEG signals.(2)The pattern recognition method of EEG signal is studied and improved.The advantages and disadvantages of OVR-SVM,OVO-SVM and DAG-SVM classification methods are analyzed,and a novel hierarchical SVM algorithm(HSVM)is proposed to classify the four classes of EEG signals.The classification validity of HSVM was verified by using the competition data set 2a in 2008.Although the training time of the proposed method is slightly longer than DAG-SVM,and the test time is not much difference between DAG-SVM,but the correct rate of classification was significantly higher than DAG-SVM.(3)The dual tree complex wavelet transform is used for feature extraction of EOG signals,solving the problems such as the frequency aliasing and shift sensitivity which ocuured in the discrete wavelet transform method.According to the Donoho threshold denoising theory,the threshold value of each layer of wavelet coefficients is estimated in DTCWT decomposition.The threshold was used for reconstruction of EOG signals.Then,the feature of EOG signal was extracted by down sampling of the reconstructed EOG signal.SVM and time domain features threshold method are used for classification.SVM classifier was applied to identify the eye movement in horizontal direction or vertical direction.Then,identification of eye movement patterns in horizontal or vertical direction was based on the threshold of the time domain feature of EOG signals.Experimental results show that these classification methods can effectively identify the saccade and continuous blink patterns of eye movement.(4)By expanding the EOG signals acquisition electrodes on the Emotiv EPOC EEG acquisition instrument,an experimental platform for the online control of hybrid brain computer interface was built,and the data was transmitted to the computer in real time.Graphical user interface for controlling the four axis aircraft was compeleted based on MATLAB2015 and VS2010 softwares.In addition,error correction mechanism is designed.The online experiment results show that the HBCI system can real time control the four axis aircraft to take off,forward,left,right,backward and landing well.