Research On Hybrid Signal Processing And Application Of P300 And SSVEP Oriented To Brain Computer Interfaces

Brain-Computer Interface(BCI)is a communication channel between a human brain and an external computer.In recent years,researchers have proposed the concept of hybrid BCI that compromises two or more types of brain signals to compensate for their respective shortcomings and increase the system's real-time and practicability.P300 potentials and Steady State Visual Evoked Potential(SSVEP)is the EEG signal commonly used in brain-computer interface(BCI).However,the long recognition cycle of P300 potentials and the finite recognizable frequency of SSVEP potentials limit the recognition rate and the number of instructions of BCI system,respectively.In this paper,a hybrid EEG BCI system based on P300 and SSVEP is designed to improve the recognition accuracy and transmission rate in the brain computer interface in view of the deficiency of the above two kinds of signals.The main research work is as follows:(1)According to the principle and characteristics of P300 and SSVEP signal,combined with the characteristics of P300 signal evoked paradigm and SSVEP signal evoked paradigm,a scheme based on P300 and SSVEP mixed EEG signal region paradigm is proposed and the corresponding mixed EEG signal visual stimulus paradigm is designed.The paradigm combines the characteristics of the number of P300 commands and the fast recognition rate of SSVEP,and replaces the first level flicker of the traditional P300 regional paradigm with SSVEP,which improves the recognition rate.(2)According to the signal processing process in the brain computer interface and the induced principle and characteristics of P300 and SSVEP signal,combined with the signal processing theory that the hybrid EEG signal should first be separated and then be parallel processed,a hybrid EEG signal processing program based on P300 and SSVEP orienting the BCI has been put forward.In the P300 signal processing,a multi-sensor weighted support vector machine algorithm(msw-SVM),reducing the difference in P300 signal which will make the amplitude difference in the target stimuli relative to non-target stimuli amplitude difference more obvious,is proposed to get better classification results.In SSVEP signal processing,a set of multivariate exponential synchronization algorithm Mset-MSI based on the multiple sets of canonical correlation analysisis proposed to improve the accuracy of signal recognition.(3)According to the research of theory above,character input system based on mixed EEG signal is exploited,the corresponding function module is designed,and the Socket communication based on UDP protocol is used to realize the data transmission between the various functional modules to conduct real time online test.At last,character input based on P300 and SSVEP signal of mixed brain computer interface is structured.And in a non-shielded environment,all the different testers successfully completed the test.