Research On Brain-Computer Interaction Method Base On AR-BCI

Patients with amyotrophic lateral sclerosis and stroke have normal brain consciousness,but due to motor dysfunction,they cannot communicate with the outside environment normally.With the development of the economy and the rise of the medical level,the basic life and material level of the disabled have been improved,but the spiritual and cultural needs of this group have been ignored.At present,although other investigators can use the brain-computer interface technology of the PC screen to allow patients to control the external environment through "ideas",the poor portability of the PC screen and the limited field of view limits the application of BCI.AR-BCI is a combination of the Virtual-real fusion of augmented reality(AR)and brain-computer interface(BCI)motion control capabilities while providing users with a portable,wide-viewing operation space,which has great potential in assisting the disabled.However,there are relatively few AR-BCI related studies,how to properly design the information display in the augmented reality space remains to be studied.Therefore,this thesis starts with designing a reasonable SSVEP stimulus layout and exploring the number of stimuli that can be accommodated,by establishing an ARMind Chess chess game system and evaluating the feasibility and practicability of the AR-SSVEP system to promote its practical process.The following is the main research work of this thesis:(1)The Augmented reality ARMind Chess chess information display was realized in this thesis.The HSG rendering engine was used to design the visual stimulus presentation method;the space anchor technology and gaze target detection technology was employed to ensure that the operator could accurately place the virtual chessboard and control the virtual chess pieces;combined with the above technology,the augmented reality chess layout information display system was constructed.(2)The EEG processing algorithm of the chess game system was designed in this paper.By analyzing the EEG processing algorithm of the mainstream SSVEP-BCI system,four experimental paradigms of different stimulation layouts were designed in the AR-SSVEP system;through the CCA algorithm,the stimulus layout with the optimal classification correct rate was reached and the effect of the stimulus layout on the classification correct rate was obtained when the stimulus targets were four;then compared the significant differences between the two in the power spectrum and recognition rate.(3)AR-BCI powered ARMind Chess prototype system was developed.The difference in amplitude,signal-to-noise ratio and classification accuracy of AR-SSVEP and PC-SSVEP was investigated under the four kinds of stimulus targets;the reasons why increasing the number of stimuli would reduce the accuracy of AR-SSVEP classification was discussed;Finally,the online chess game of ARMind Chess prototype system was established,and the man-machine game mode of virtual reality and man-person game mode of augmented reality were designed.The online experiment results of 5 participants demonstrate that the chess game of ARMind Chess is feasible and practical,and is expected to enrich the spiritual life of stroke patients in the future.