Research On Steady State Visual Evoked Potential Based Brain-Computer Interface In Argumented Reality

Brain-Computer Interface(BCI)has developed rapidly during the last decades.Especially in steady state visual evoked potential(SSVEP)based BCIs,which has achieved high information transfer rate(ITR).However,high-performance noninvasive BCIs usually need display devices to present the visual stimulus to evoke specific EEG patterns.The most popular display device now is computer monitor which is not portable thus restrict the portability of BCI.Besides,visual evoked BCIs needs subject to concentrate on special visual stimulus on the monitor,which restrict the subject’s ability to acquire surrounding environment information.This problem led to uncomfortable communication between human and machine.By combining argument reality(AR)technology with BCI,we could resolve above problems and achieve a more practically applicable BCI system.However,up to now,there are only a few research about AR-BCI.The accuracy and ITR of AR-BCI remains to be improved.This study proposed an AR-BCI system based on Microsoft Hololens AR device and lab made electroencephalograph(EEG)amplifier.We investigated the impact of algorithm on accuracy.14 subjects took part in the offline experiment.Results show that Extended-Filter Bank Canonical Correlation Analysis has the highest accuracy.With 1s,2 s,3 s EEG data,extended-FBCCA can achieve 87.7%,85.4%,97.6% accuracy on average,the corresponded ITR is 64.6 bit/min,62.9 bit/min,55.6 bit/min..Further in this study,we build an AR-BCI robotic arm control system.With the system we investigate the impact of background environment on the performance of AR-BCI system.10 subjects took part in the experiment and achieved an average accuracy of 81.2%,and average online ITR of 35.8 bit/min.This result shows that complicated dynamic environment has negative effect on performance of AR-BCI.Eventhough,we compared the performance of our AR-BCI system with related researches and find out that our system has realized better performance than other ARBCI systems.Finally,this study proposed a consecutive SSVEP visual stimulus,timing synchronize and target detection method.We test the feasibility of consecutive SSVEP method on screen based BCI.Then we combined the AR-BCI with consecutive SSVEP to control an air drone.We compare the performance of consecutive SSVEP AR-BCI with regular AR-BCI in air drone control.5 subjects took part in this experiment and achieved ITR of 49.1 bit/min and out put 27 correct command per minut in consecutive SSVEP AR-BCI while regular AR-BCI achieved ITR of 41.2 bit/min and out put 18.7 correct command per minut.The result shows that consecutive SSVEP method can increase the performance of AR-BCI system.This study investigate the influencing factor of AR-BCI performance and methods to enhance AR-BCI performance.The result of this study laid foundation for realizing portable and comfortable high performance BCI.