The Research And Implementation Of The Hybrid Brain-computer Interface System

Brain-Computer Interface(BCI),as a new human-computer interaction technology,can establish a direct information interaction channel between the brain and the external devices rather than relying on the peripheral nerve and muscle channels.BCI technology has a broad application prospect in the fields of medical rehabilitation,game entertainment,military and other fields.In recent years,with the development of signal processing and computer technology,the research in BCI has achieved great achievements.However,there are still some problems to be solved in commercial application,such as the system recognition accuracy and asynchronous control etc.,Therefore,The construction of practical online BCI system is of great significance for the development of BCI technology.Independent Component Analysis(ICA)and Common Spatial Pattern(CSP),as the most widely used electroencephalogram(EEG)processing technology,have their own advantages and disadvantages in BCI.In view of the characteristics of unsupervised learning of ICA,it is used as the feature extraction method of BCI system in this thesis,and on this basis,a zero training classifier for three-class motor imageries is constructed.Meanwhile,the subject's blinking behavior is identified by the analysis of Electrooculogram(EOG),and the subjects autonomously control the switch between the task state and the non task state by blinking so as to realize the asynchronous control of the BCI system.Eventually,the online hybrid MI-BCI system was developed under the VC++ platform.The specific work of this thesis are as follows:1.The Event-Ralated Desynchronization/Synchronization(ERD/ERS)phenomena related to motor imagery EEG was studied.The feature extraction methods such as time domain,frequency domain and spatial domain filtering of EEG signal were introduced,and the typical pattern classification algorithms were described.2.A optimization method of EEG training samples based on CSP was designed to eliminate low quality training data.Experimental results showed that this method can effectively improve the performance of CSP spatial filter.Then the ICA algorithm used in MI-BCI was introduced,and it compared with CSP in many aspects to draw the conclusion that the ICA algorithm is more suitable for the application in online MI-BCI.3.A new hybrid MI-BCI system based on EEG and EOG was designed.The system provides a simple human-computer interface and GUI control interface.It has the characteristics of fast data transmission,strong real-time and good operability,and the parameters for different users can be reset.The subjects determine the beginning of the motor imagery by blinking,and the three-class motor imageries are used to achieve the movement control of the target in GUI.4.The experimental paradigm of online MI-BCI system was implemented,and the offline analysis function was provided.Four subjects participated in online experiments.The experimental results showed that participants could effectively control the target in GUI to reach the destination though motor imagery,and the motion trajectory of subjects were highly matched with the planed path.This system was developed on VC++ platform with the characteristics of convenient operation,simple use,low complexity of algorithm and high practical value.