Study On Brain-computer Interface Of Visual Evoked Potentials Based On Pseudo-random Code Modulation

Brain-computer interface(BCI)does not relying on the muscles and peripheral nerves and can achieve communication with the outside world,is a direct connection through the brain to achieve real-time external communication system.Nowadays,with the rapid development of brain-computer interface technology,researchers have done a lot of research on Steady-state visual evoked potential BCI(SSVEP),while Pseudo-random code modulated visual evoked potential BCI(c-VEP BCI)is relatively few,the traditional c-VEP BCI is to use one or a coding and its time shift to modulate different stimulus targets,limiting the number of targets increases,thus limiting the BCI system information transfer rate.In this paper,the c-VEP BCI is studied from the method of signal processing and improving the number of stimulus targets of c-VEP BCI.The visual evoked potential brain-computer interface based on a pseudo-random sequence modulation is to optimize the airspace filter by Canonical correlation analysis(CCA),and use the template matching method(TMM)for target recognition.In this paper,we propose a method to optimize the airspace filter by Signal Fraction Analysis(SFA),and classify and identify targets using one class support vector machines(OCSVM).The two methods of spatial filtering and two kinds of classification and recognition methods can be combined with each other,Four different forms of the method,the experimental results show that these four methods have achieved a high recognition accuracy.In this paper,we propose a visual evoked potentials brain-computer interface based on modulation of more than a pseudo-random sequences,which are used to group the stimulus targets by two different Golay codes and a near perfect sequence,which achieves a stimulus with 48 targets Device,a substantial increase in the number of stimulating targets.the airspace filter is optimized by Canonical correlation analysis,and the template matching method is used to classify and identify the target,and the classification and recognition rate is vey high.The system uses 8 experimental data of the subjects were analyzed,the classification accuracy rate reached 94.95%.