实时脑—机接口设计与研究

发布时间：2018-06-26 18:11

本文选题：脑-机接口 + 脑电　；参考：《重庆大学》2012年硕士论文

【摘要】：脑-机接口（Brain-Computer Interface，BCI）技术形成于20世纪70年代，近年来在国际上成为一个热点并获得了快速的发展。它是一种不依赖于大脑的正常输出通路（肌肉组织和脑外周神经系统），直接在人脑与计算机或其他设备之间进行信息交换的通讯系统。脑-机接口为那些有严重运动障碍但脑神经功能正常的患者提供了一种与外界交流的途径，提高其生活质量。此外在工业、航空、军事等领域也有潜在的应用价值。针对实时脑-机接口对系统的通讯速率、信号处理速度、准确率要求很高的问题，本文对实时脑-机接口的设计进行了研究。本文以LabVIEW作为编程语言，建立了脑电信号的实时传输和处理系统；通过对TCP连接的数据包格式、封装协议和连接过程的研究，设计了接口电路及通信协议，在此基础上设计并建立了基于脑电α波的实时光标控制系统和基于眼电的实时遥控汽车控制系统，受试者的操作实验结果表明系统实时性强，准确率高。脑电信号的特征提取和模式识别是脑-机接口系统研究中的重点。本文首先利用BCI2003国际竞赛公开的左右手运动想象数据进行了离线分析，通过对C3、C4导联脑电信号的经典功率谱分析确定了以μ节律作为特征提取的频段；再利用AR模型分析受试者左右手运动想象脑电信号，对比不同阶数下的频谱图，得出12阶AR模型能够取得良好的分辨率，由此确定用μ波平均能量和12阶AR模型系数作为特征向量，采用Fisher线性判别分析作为模式识别分类器。然后自主设计同步模式在线实验进行特征提取和模式识别，，实验结果也取得了较高的识别正确率，为下一步进行了基于左右手运动想象的异步模式脑-机接口研究提供了理论和实验的支持。本文通过对脑电信号的离线分析，确定了实时脑-机接口系统设计的特征提取方法；建立了基于脑电α波和眼电信号的实时脑-机接口系统，为本项目组将来的脑-机接口研究建立了实验平台；基于运动想象的同步模式在线脑-机接口的实现为基于运动想象的实时脑-机接口研究积累了经验。
[Abstract]:Brain-Computer Interface (Brain-Computer Interface) (Brain-Computer Interface) (Brain-Computer Interface) (BCI) technology was formed in 1970s. It is a normal output pathway independent of the brain (muscle tissue and peripheral nervous system) that exchanges information directly between the human brain and computers or other devices. Brain-computer interface (BCI) provides a way to communicate with the outside world and improve the quality of life for those patients with severe motor disorders but normal brain nerve function. In addition, there are potential applications in industry, aviation, military and other fields. Aiming at the problem that the communication rate, signal processing speed and accuracy rate of the real-time brain-computer interface are very high, the design of the real-time brain-computer interface is studied in this paper. In this paper, we use LabVIEW as the programming language to establish the real-time transmission and processing system of EEG signal, and design the interface circuit and communication protocol by studying the data packet format, encapsulation protocol and connection process of TCP connection. On this basis, a real time cursor control system based on 伪 wave of EEG and a real time remote control system of vehicle based on eye electricity are designed and established. The experimental results show that the system is real-time and accurate. The feature extraction and pattern recognition of EEG signal are the key points in the study of BCI system. In this paper, off-line analysis of left and right hand motion imagination data from BCI2003 international competition is carried out, and the frequency band based on 渭 rhythm is determined by classical power spectrum analysis of C3C4 lead EEG signal. Then using AR model to analyze the subjects' left and right hand motion imagination EEG, comparing the spectrum of different order, it is concluded that the 12 order AR model can obtain good resolution. The 渭 wave average energy and 12 order AR model coefficients are used as eigenvector and Fisher linear discriminant analysis is used as pattern recognition classifier. Then we design a synchronous pattern online experiment for feature extraction and pattern recognition. The experimental results also achieve a high recognition accuracy. It provides theoretical and experimental support for the further research of asynchronous brain-computer interface based on the left-right-hand motion imagination. Based on the off-line analysis of EEG signal, the feature extraction method of real-time brain-computer interface system is determined, and the real-time brain-computer interface system based on EEG alpha wave and Eye-electric signal is established. The experiment platform is established for the future brain-computer interface research of the project team, and the realization of online brain-computer interface based on motion imagination has accumulated experience for real-time brain-computer interface research based on motion imagination.
【学位授予单位】：重庆大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP334.7

【相似文献】