基于嵌入式的脑机接口系统设计

发布时间：2018-06-29 17:06

本文选题：脑机接口 + 嵌入式　；参考：《电子科技大学》2017年硕士论文

【摘要】：脑机接口(Brain Computer Interface,BCI)系统是一种不依赖脑正常输出通路就可以实现对脑外事物进行控制的系统,由于其具有巨大的应用前景以及理论研究价值,近年来,其已成为脑科学、生物医学工程及人机自动控制领域的一个研究热点。脑机接口发展到现阶段,已经有了很大进步,但是,目前在研的BCI系统大多依赖于基于PC平台的高性能处理器,PC体积较大,便携性差,给脑机接口的实际应用带来了一定的限制。便携性是脑机接口亟待解决的问题,近年来,ARM等嵌入式处理器有了较大的突破,其运算能力和图像处理能力都有了质的飞跃,使用嵌入式处理器已经可以应对一些较为复杂的运算任务,基于此,本文研究设计了一种基于嵌入式平台的脑机接口系统,系统采用ARM架构的嵌入式处理器作为主控,搭载嵌入式Linux系统,实现了4个指令的控制,且指令数还可以在后续改进版本中进一步扩展,该系统体积小,方便携带,也可以和电控轮椅等外围设备配合使用,从而增强其实用性。本文的主要内容包括嵌入式系统搭建、USB驱动程序设计以及脑机接口应用程序设计三个部分。其中嵌入式系统搭建包括了Linux内核裁剪、文件系统制作及QT库安装三个部分;USB驱动实现的是脑电放大器到嵌入式系统的单向数据传输;脑机接口应用程序包含了波形显示程序和脑机接口程序两部分,整个程序实现了视觉刺激、脑电数据采集、特征提取以及模式分类等内容。最后,为了验证该嵌入式脑机接口系统的性能,对该系统的核心模块以及系统整体性能进行了测试,测得该脑机接口系统的平均准确率为83%,平均信息传输率为20比特每分钟,测试结果表明,该系统能实现基本的控制任务,满足基本的应用需求。
[Abstract]:Brain computer Interface (BCI) system is a kind of system which can control the external brain without relying on the normal brain output pathway. Because of its great application prospect and theoretical research value, it has become a brain science in recent years. Biomedical engineering and man-machine automatic control field is a research hotspot. There has been great progress in the development of brain-computer interface at this stage. However, most of the BCI systems studied at present rely on high performance processors based on PC platform, such as large size and poor portability. It brings some limitations to the practical application of BCI. Portability is an urgent problem in brain-computer interface. In recent years, arm and other embedded processors have made a great breakthrough, their computing power and image processing ability have made a qualitative leap forward. Using embedded processor can deal with some complicated computing tasks. Based on this, a kind of brain-computer interface system based on embedded platform is designed in this paper. The embedded processor based on arm architecture is used as the main control in the system. With embedded Linux system, the control of four instructions is realized, and the number of instructions can be further expanded in the subsequent improved version. The system is small, portable, and can be used in conjunction with peripheral devices such as electrically controlled wheelchairs, etc. So as to enhance its practicability. The main contents of this paper include three parts: the design of USB driver for embedded system and the design of BCI application program. The embedded system consists of Linux kernel clipping, file system fabrication and QT library installation. The USB driver realizes the one-way data transmission from EEG amplifier to embedded system. The BCI application program includes two parts: waveform display program and brain-computer interface program. The whole program realizes visual stimulation, EEG data acquisition, feature extraction and pattern classification. Finally, in order to verify the performance of the embedded BCI system, the core modules of the system and the overall performance of the system are tested. The average accuracy of the BCI system is 833 and the average information transmission rate is 20 bits per minute. The test results show that the system can achieve basic control tasks and meet the basic application requirements.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R318;TP316

【参考文献】