Research On The Construction Method And Key Technology Of Portable Brain-Computer Interface Based On SSVEP

As a new type of human-computer interaction way,brain-computer interface(BCI)can directly translate brain signals into machine-readable operating instructions without the involvement of peripheral nervous systems and muscle tissues,which expands people's ability of information communication and control.Generated by the cerebral visual cortex,the steady-state visual evoked potential(SSVEP)is an inherent response to the periodic visual stimuli.With the advantage of fast information transfer rate,high classification accuracy and without the need of training,SSVEP-based brain-computer interface systems have a promising future to be used in daily-life.However,most of the current SSVEP-BCIs employ expensive electroencephalography(EEG)recording devices and bulky visual stimulators,limiting their use in dailylife.Therefore,how to realize portable SSVEP-BCI systems,with good mobility,high practicability and low cost,has become a popular topic and a difficult problem in the current SSVEP-BCI research.To address this issue,his paper presented a systematic research on developing a portable brain-computer interfacing system based on SSVEP,which employed a wearable and wireless Emotiv EPOC headset as an EEG recording device and a mobile phone as a tool to present visual stimuli.Firstly,considering the disadvantage of the traditional stimulus flickering method,which is based on frame-by-frame flicker way and is low user-friendly,a new stimulus flickering method,which was based on the changes in gray values of the stimuli,was proposed.This method produces periodic visual stimuli by sinusoidally varying the gray value.The experiment results show that this stimulus modality is less irritating to eyes and can induce SSVEP signals with higher quality.Secondly,in order to obtain a higher classification accuracy of SSVEP signals,three improved SSVEP classification algorithms based on canonical correlation analysis(CCA)are presented,i.e.,the frequency distance-based CCA,the fuzzy set-based CCA,and the individual signal template-based CCA.Testing results indicate that the classification effects of these three algorithms are enhanced in turn.Finally,on the basis of the research described above,a brain-controlled portable smart-car system has been implemented.The system sends a control command after dealing with two epochs of SSVEP signals,and can produce a variety of control commands with few types of stimulus frequencies,showing an advantage for practical applications.The experiments show that the proposed system obtain the command recognition accuracy and the information transfer rate of the system are 79.45%10.92 bits/min respectively.