Design And Research Of Brain-Computer Interface System Based On SSVEP

As a new communication means of human computer interaction, brain-computerinterface (BCI) transforms electroencephalogram into the commands of computercontrol, which is independent of peripheralnervous and muscle. Nowadays, with thedevelopment of BCI, researchers have paid more attention to practical applicationother than theoreticalaspect. With a goal of practicability of brain-computer interface,a BCI platform based on steady-state visual evoked potentials (SSVEP) is developedin this dissertation, on which a calling system and a cursor control system arerealized.The BCI system based on SSVEP includes visual stimulator, signal acquisitionmodule, real-time signal processing module. The canonical correlation analysis (CCA)is employed to recognize targets by searching the maximum correlation valuebetween SSVEP signal and reference signal. Compared with the power spectraldensity analysis (PSDA) method, the online experiment result indicates that the CCAmethod can improve the accuracy rate and information transfer rate of the system.In calling system, the SSVEP is transformed into control command firstly andsent to the SIM300module to accomplish the calling operation. With the CCAmethod, the average accuracy rate and information transfer rate reach up to93.16%and33.53bit/min, respectively, which indicates that the system has high classificationaccuracy and information transfer rate. In cursor control system, the SSVEP istransformed into control command, and then transferred to control the cursor on GUI.The average accuracy rates of single task and dual-task are90%and87.5%,respectively, which indicates that the system has high recognition performance andcan satisfy the demand of the disabled as well.