Research And Application Of Virtual Man Control Based On Brain Computer Interface

The researches of Brain-computer interface(BCI)technology have always been focused on patients with movement disorders but normal brain function.Although many BCI systems have been created in the past decades and have formed a basic composition framework of BCI,it has many problems to solve for the application of BCI because of the limitation of BCI technology itself and other factors.In order to improve the discriminant efficiency and broaden the range of application of BCI,virtual reality(VR)technology is introduced to BCI system,and use BCI as a new input mode to control VR systems.the main contributions of this paper could be concluded as below:First,it collated and analyzed the relative researches about components of EEG and current BCI systems,and found that P300 of ERP is highlight in the process of decision making.It designed and created a P300 based BCI training platform.In this platform,a 3D stimulate sequence is employed to evoke P300,and it has a better trigger effect than traditional 2D stimulate sequence.Bayesian linear discriminant analysis(BLDA)is used in this system to classify the P300 feature signals and gets a good effect from the training data of 8 subjects.,and have some improvement for previous BCI training system in some extend.Second,the relative knowledge of computer graphics is collated and used to create the human-computer interaction system.In this paper,it first realized a VR system by OSG,and subjects could use keyboard to control a virtual man interacting with virtual scene.This system have features of high flexibility,strong expansibility and could well support hardware devices,and gets a good man-machine interaction effect after tested by keyboard and make a solid foundation for BCI application expanding.Third,as the application range of current BCIs is small,and the input mode of current VR is rigid and simple,it first realized a P300 based virtual man obstacle avoidance system and a self-adaptive algorithm is applied in this system to make the virtual man control system more fluent.In this system,3 subjects are tested to could control virtual man with their brain to move in the virtual scene for 4 times,their task finishing time,recognition accuracy,collision times demostrated that the system is feasible.