基于虚拟现实的运动想象脑—机接口训练系统

发布时间：2018-01-15 14:13

本文关键词：基于虚拟现实的运动想象脑—机接口训练系统　出处：《郑州大学》2017年硕士论文　论文类型：学位论文

【摘要】：脑-机接口(Brain-Computer Interface,BCI)是在大脑和计算机及外部设备之间由人工搭建的一条不同于传统大脑信息传输的通路,作为一种全新的信息交流和控制技术,可使用户直接通过意识与外部设备实现信息交互。运动想象(Motor Imagery,MI)脑-机接口作为一种自主BCI方式,在神经退行性疾病患者的康复、军事远程作战等方面具有巨大的潜在价值,成为近年来的研究热点。近十几年来,运动想象脑-机接口的研究取得较大进展,但依然存在以下问题:(1)受试者运动想象能力差异较大,不同受试者对同一个BCI系统的操控能力具有很大差异;(2)不同受试者做同一思维任务时脑电信号的特征不尽相同,需要选择最优特征提取算法;(3)不同受试者在做同一思维任务时大脑的活动区域不尽相同,需要选择不同的脑电导联;(4)训练对受试者的运动想象能力具有较大的影响,需要对受试者进行大量训练,以提高受试者的操控能力。针对以上问题,本文把虚拟现实(virtual reality,VR)和脑-机接口技术相结合,设计了一款自适应运动想象BCI系统(Motor Imagery Brain-Computer Interface,MI-BCI)的虚拟现实训练与实时控制系统。本文的主要工作和创新点如下:(1)设计了一款自适应算法选择的运动想象BCI软件系统。本文利用C#和MATLAB混合编程方式,设计实现了MI-BCI软件。集成了三种脑电信号预处理算法、五种特征提取算法和两种分类识别算法,提供离线分析、在线分析和自适应算法选择功能。根据受试者的脑电信息,自动选择最适合该受试者最优算法组合,增强了BCI系统与受试者的契合度,提高了BCI系统的普适性。(2)提出了一种新的导联选择算法——遗传蜂群算法。针对不同受试者大脑活动区域差异的自然现象,本文利用遗传算子改进传统蜂群算法邻域搜索操作,提出了一种基于遗传算子的人工蜂群算法(artificial bee colony algorithm based on genetic operators,GA-ABC)对运动想象BCI系统进行导联优选,大大降低了导联维数,提高了脑电信号处理速度和识别率,为BCI系统步入实用提供了参考。(3)搭建了一个基于Unity3D的运动想象虚拟现实场景。采用Unity3D、DAZ Studio和3Dmax设计了虚拟场景和3D人物模型与动作,通过Socket套接字实现了BCI系统与虚拟现实系统的信息交互。虚拟现实场景功能分为受试者训练和训练效果测试,其中测试部分又分为同步在线测试和漫游测试,分别提供第一视角和第三视角两种模式供用户选择,为受试者的运动想象训练和实时操控提供了一个直观的反馈,提高了受试者的运动想象能力和对BCI系统的操控能力。(4)设计了运动想象脑-机接口虚拟现实在线控制实验。本文设计同步在线实验和虚拟现实在线漫游测试两种实验对四名受试者的训练效果进行测试,根据实验结果得出本文设计训练系统能够提高受试者运动想象能力,同时验证了本系统设计是有效性、合理性和可行性。
[Abstract]:Brain computer interface (Brain-Computer Interface BCI) is a different from the traditional brain information transmission built by artificial pathway between brain and computer and peripheral equipment, as a new information communication and control technology, users can directly through the implementation of information interaction consciousness and external equipment. Motor imagery (Motor Imagery MI) brain computer interface as an independent BCI, in patients with neurodegenerative disease rehabilitation, has great potential for military remote operations and so on, become a research hotspot in recent years. In recent years, the movement of imagination BCI has made great progress, but there are still the following problems: (1) the subjects of motor imagery ability is different, has different ability to control the same BCI system in different subjects; (2) different subjects have the same mental task EEG signal characteristics vary At the same time, to select the optimal feature extraction algorithm; (3) the different subjects in the same region of the brain thinking activities are not the same, need to choose different EEG lead; (4) exercise training on the subjects of the imagination has remarkable effect, the need for a large number of training subjects and in order to improve the subjects' ability to control. To solve the above problems, the virtual reality (virtual reality, VR) and brain computer interface technology, design a adaptive motion system (Motor Imagery Brain-Computer BCI thought Interface, MI-BCI) of the virtual reality training and real-time control system. The main work and the innovation points are as follows: (1) designed to choose a adaptive algorithm of motor imagery BCI software. Using mixed programming of C# and MATLAB, the design and implementation of MI-BCI software. The integration of the three kinds of EEG preprocessing algorithm, five. Feature extraction algorithms and two kinds of classification algorithms, provide off-line analysis, on-line analysis and adaptive algorithm selection. According to the EEG information of subjects, automatically select the most suitable for the subjects of optimal algorithm combination, fit BCI system and enhance the subjects, improve the BCI system universality. (2) proposed a new channel selection algorithm -- genetic colony algorithm. For different natural phenomena subjects brain activity of regional differences, this paper uses the improved genetic operators traditional colony algorithm of neighborhood search operation, put forward a kind of artificial bee colony algorithm based on genetic operator (artificial bee colony algorithm based on genetic operators, GA-ABC to optimize the motor imagery) lead BCI system, greatly reducing the lead dimension, improve the EEG signal processing speed and recognition rate, providing a reference for the BCI system into practical. (3) set up Imagine a virtual reality scene based on the movement of Unity3D by Unity3D. DAZ, Studio and 3Dmax to design the virtual scene and the 3D character models and action, through the Socket socket to realize the information interaction between BCI system and virtual reality system. The virtual reality scene is divided into subjects of training and training effect test, the test part divided into synchronous online testing and roaming test, provide a first view and third view two modes for user selection, training and real-time control provides an intuitive feedback for subjects with motor imagery, improve the subjects' movement imagination and the ability to control the system of BCI (4) is designed. Motor imagery BCI Online virtual reality control experiment. This paper design of synchronous online experiment and virtual reality roaming online testing and training effect of two kinds of experiments on four subjects tested, according to the The experimental results show that the design training system can improve the ability of the subjects to imagine motion, and the validity, rationality and feasibility of the design of the system is verified.

【学位授予单位】：郑州大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R318;TP391.9

【相似文献】