Research And Application Of Online Real-time Brain-computer Interface System Based On Motor Imagery

Brain-Computer Interface(BCI)is a new type of human-computer interaction technology that can bypass peripheral nerves and muscles and directly use brain electrical signals to control external devices.It involves brain science,information technology,and control technology.Subjects have been applied to various fields such as military,medical,and entertainment.Brain-machine interface based on sports imaging brain electricity is a very important paradigm.Its performance is quite remarkable,and it can build a communication with the outside world for patients with nerve injury.aisle.In this paper,motion Imagining EEG is the object of study.It elaborates the mechanism of EEG generation in sports imagination and introduces the method of EEG acquisition in sports imagination.However,for the characteristics of nonlinear,non-stationary,and easily disturbed,the left and right hand movements are designed.Imagining the brainwave signal acquisition scheme,we successfully obtained five subjects' EEG data and then preprocessed it.By comparing the classification effects of different adaptive preprocessing algorithms,the processing method corresponding to the optimal classification results is selected,namely the combination of Fast Independent Component Analysis(FastICA)and wavelet packet decomposition.Feature extraction is a key part of the BCI system.In this paper,5 subjects were trained to perform left and right hand motion imaging EEG features using the AR model and the common space model(CSP).Support vector machine(SVM)was used as a classifier for classification.It shows that the CSP feature extraction method has a higher classification accuracy than the AR model.In the pattern classification process,the influence of the error penalty factor C and the nuclear parameter ? in the SVM classification on the final classification result was discussed in detail,and its optimal value was successfully found.Finally,on the basis of analyzing the preprocessing,feature extraction and pattern classification algorithms,an online real-time brain-computer interface system based on sports imagination EEG was successfully built,connected with the brain control robot testing platform provided by the second national brain-computer interface contest.Simulating the control results in the actualenvironment,all five subjects successfully avoided obstacles within a certain period of time and reached the end successfully.The reliability of the system was verified.The establishment of this system can provide a certain theory for the brain-computer interface out of the laboratory.basis.