Brain Computer Interface Training System Based On Virtual Reality

Brain computer interface(BCI)is an artificial way that is different from the traditional brain information transmission path,which is established between the brain and the computer and external devices.Brain computer interface,as a new technology of information exchange and control,enables users to communicate with each other directly by means of consciousness.Motor Imagery as a kind of autonomous BCI,has great potential value in the rehabilitation of patients with neurodegenerative diseases,military remote operations and so on,and has become a hot research topic in recent years.In recent years,great progress has been made in the research of motor imagery brain computer interface,but there are still the following problems:(1)there were significant differences in the ability of the subjects' imagination in the same BCI system,and different subjects had great differences in the control ability of the same BCI system;(2)different subjects do the same thinking task when the characteristics of EEG signal are not the same,need to select the optimal feature extraction algorithm;(3)different subjects have different brain activation regions when they are doing the same task,need to choose different EEG channels;(4)the training has a great influence on the subjects' motor imagery ability,so they need to do a lot of training to improve the control ability of the subjects.To solve the above problems,this thesis combines the virtual reality and the brain computer interface technology,and designs an adaptive motion imagine BCI virtual reality training and real-time control system.The main work and innovations of this thesis are as follows:(1)An adaptive algorithm named motion imagination BCI software system was designed.This thesis designs the MI BCI software using C# and MATLAB.The system integrates the three kinds of EEG preprocessing algorithm,five kinds of feature extraction algorithm and two kinds of classification algorithm,providing offline,online analysis and adaptive algorithm selection function.According to the EEG information of the subjects,the best combination of the optimal algorithm is selected automatically,which enhances the BCI system fit with the subjects and improves the universality of the BCI system.(2)A new algorithm of lead selection was proposed,which is the artificial bee colony algorithm based on genetic operator.According to the phenomenon of different active regions of brain in different subjects,this thesis presents an artificial bee colony algorithm based on genetic operator for MI-BCI system to choose the best channel combination for different subjects.The algorithm can reduce the lead dimension and improve the EEG signal processing speed and recognition rate,which provides a reference for BCI system in the practical.(3)A virtual reality scene of motion imagination based on Unity3 D engine was constructesd.The virtual scene and 3D character model and action are designed by using Unity3 D,DAZ Studio and 3Dmax,and the information interaction between BCI system and virtual reality system is realized by Socket.Virtual reality scene function is divided into the subject training and training effect test,in which the test part is divided into synchronous online test and roaming test,providing the first view and the third perspective of the two modes for the user to choose.The system provides an intuitive feedback for the subject's motion imagination training and real-time manipulation,improving the athlete's ability to visualize and manipulate the BCI system.(4)The virtual reality on-line control experiments of motion imagination brain-machine interface were designed.In this thesis,two kinds of experiments,synchronous online experiment and virtual reality online walkthrough test,are designed to test the training effect of four subjects.According to the experimental results,it is concluded that the training system designed in this thesis can improve the motor imagination ability of the subjects.At the same time,it is proved that the system design is effective,reasonable and feasible.