基于多模式EEG的脑-机接口虚拟键鼠系统设计  

A Hybrid BCI( Brain-Computer Interface) Based on Multi-Mode EEG for Words Typing and Mouse Control

作　　者：谢松云[1] 刘畅[1] 吴悠[1] 张娟丽[1] 段绪[1] 

机构地区：[1]西北工业大学电子信息学院,陕西西安710072

出　　处：《西北工业大学学报》2016年第2期245-249,共5页Journal of Northwestern Polytechnical University

基　　金：国家自然科学基金(61273250);陕西省工业攻关项目(2015GY003);西北工业大学研究生创业种子基金(Z2015112)资助

摘　　要：现有的脑-机接口系统大都只基于单模式的脑电特征,系统能实现的功能非常有限,从而制约了脑-机接口系统的应用。采用基于多种模式脑电信号(electroencephalogram,EEG)的脑-机接口技术来实现虚拟键鼠系统,使得被试可以利用自身的脑电信号控制鼠标和键盘的操作。研究了脑-机接口中常用的3种脑电信号,分别是P300波、alpha波以及稳态视觉诱发电位(steady state visual evoked potential,SSVEP),通过设计实验成功的诱发出了被试相应的特征脑电信号。利用SSVEP的脑电特征设计6频率LED闪烁刺激的虚拟鼠标系统,实现控制鼠标光标移动、单击左键和单击右键的任务;利用P300波的脑电特征设计6×6的字符矩阵虚拟键盘系统,实现字符输入的任务;利用被试自主闭眼增强alpha波的脑电特征,实现鼠标和键盘应用切换的任务。研究了适宜这3种脑电特征的最佳测量电极组合及模式识别算法,使得对3种脑电信号的识别正确率均达到了85%以上。测试结果显示,文中设计的基于多模式EEG的脑-机接口虚拟键鼠系统能有效地实现鼠标控制以及键盘输入的任务。The existing BCI systems are mostly based on single EEG（ Electroencephalogram） feature; thus,the functions of these systems are very limited. A hybrid BCI system based on multi-mode EEG for words typing and mouse control has been designed in this paper. This paper studies three commonly used EEG features in BCIs,namely,P300,alpha waves and SSVEP. Three experiments are designed using software E-Prime to evoke the features. According to the different evoking experiments and signal processing methods,SSVEP is used to design 6flashing LED virtual mouse to move the mouse cursor and click left key and right-click,P300 is used to design a virtual keyboard of 6x6 character matrix to input the characters. The switch of the mouse and keyboard application is controlled by the enhancement of the alpha wave through closing the eyes. The real-time processing method of the three EEG features that include data segment,signal preprocessing,feature extraction and pattern recognition has also been studied; this study makes the average accuracies of the recognition of these three EEGs attain more than85%. Test results and their analysis showed preliminarily that the BCI system designed in this paper can effectively implement the mouse and keyboard input tasks.

关 键 词：脑电信号 脑-机接口 虚拟键/鼠系统 机器学习 模式识别 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]