基于交叉频率扫描技术的脑机接口设计与实现

发布时间：2018-01-09 20:31

本文关键词：基于交叉频率扫描技术的脑机接口设计与实现　出处：《电子科技大学》2015年硕士论文　论文类型：学位论文

【摘要】：脑-机接口(Brain Computer Interface,BCI)技术是不依赖于常规脑外周神经和肌肉系统,连接人脑与外界(计算机或其它外部设备)进行信息交流和控制的全新通道。脑-机接口将人脑的信号直接转换成外部设备的控制命令,信息的传递不再需要经过外周神经和肌肉等传出通道。BCI的最终目标,是让患有丧失运动能力疾病的病人,能够操作计算机、语音合成器、辅助性的工具和神经修复等设备。这样的接口可以增加患者的独立行动能力,从而提高患者生活质量,同时减轻社会的负担。本文为了解决目前脑-机接口实验中频率点资源匮乏的问题,提出了双频扫描的思路,即在一个选择目标里用两个频率进行交叉扫描刺激来获得稳态视觉诱发电位,本方法能大幅度提高频率资源的利用率,且双频扫描产生的次生频率可以提高系统的目标识别准确度,从而有可能构建复杂脑-机接口系统。根据需求我们设计并实现了基于ARM的4*4的双频扫描刺激器矩阵,并对刺激器频率点的选择与搭配进行了探讨,最后实现了刺激器通过串口与PC进行通信,并编写了使用方便的串口通讯操作界面。本文还介绍了如何对提取的脑电信号进行处理,使之成为对我们有用的信号,鉴于特征提取效率的极端重要性,作者实现了本系统的特征提取算法,先用DAUB4滤波算法对脑电信号进行预处理,再用傅里叶变换算法进行特征频率提取,并在论文中有详细描述,尤其是对傅里叶变换算法,本文详细地分析了蝴蝶操作,二进制反转,三角递归等步骤从数学公式到算法代码的产生过程。在脑电实验中,脑电信号的数据量比较大且数据产生速度快,为了准确无误的采集并处理这些数据,本文还解决了包括缓存区的管理,数据的保序,以及处理时各线程间对数据的互斥访问等问题。最后本文设计了脑-机接口实验,邀请了10名受试者通过本脑-机系统输入一个30位的字符串,通过对10名受试者的正确率的统计,发现大部分受试者的正确率都在90%以上,总的正确率也达到了88.6%,验证了脑-机系统的设计以及脑-机接口实验的设计是合理的。
[Abstract]:Brain computer interface (Brain Computer Interface BCI) technology is not dependent on conventional brain peripheral nerve and muscle system, connected with the outside world (human computer or other devices) a new channel for information communication and control. The brain machine interface will be the brain's signal is directly converted into control commands for external devices, information no longer need to transfer through the ultimate goal of the peripheral nerves and muscles from the.BCI channel, is suffering from loss of athletic ability let patients, can operate the computer, voice synthesizer, tools and other equipment auxiliary nerve repair. This interface can increase the patient's ability to act independently, so as to improve the quality of life of patients, and reduce the social the burden. In order to solve the current shortage of the frequency of brain computer interface experiment resource, put forward the idea that using dual frequency scanning, two frequency in a selected target. The rate of cross scanning stimulation to obtain the steady-state visual evoked potential, this method can greatly improve the utilization rate of frequency resources, secondary frequency and double frequency scanning the target recognition system can improve the accuracy, thus it is possible to build complex BCI system. According to the requirements we design and implement a dual frequency scanning stimulator matrix ARM based on 4*4, and the stimulator frequency point selection and collocation are discussed, finally realizes the stimulator via serial communication with PC, and the preparation of the easy to use serial communication interface. This paper also introduces how to extract the EEG signal processing, signal to be useful to us. In view of the extreme importance of feature extraction efficiency, the author realizes the system feature extraction algorithm, the first EEG preprocessing with DAUB4 filtering algorithm, and then Fu Liye transform The characteristic frequency extraction algorithm, and are described in detail in the paper, especially for the Fu Liye transform algorithm, this paper presents a detailed analysis of butterfly operation, binary inversion process steps from triangle recursive mathematical formula to the algorithm code. In the EEG, EEG signals of large amount of data and the data generation rate soon, in order to accurately collect and process these data, this paper also solves the problem of including buffer management, order preserving data, inter thread data processing and exclusive access problems. This paper design a brain computer interface experiment, 10 participants were invited to enter a 30 the string through the brain computer system, through the statistical accuracy of 10 subjects, found that the correct rate of most of the subjects are more than 90%, the overall accuracy rate reached 88.6%, to verify the design of brain computer system and brain The design of the experiment of the machine interface is reasonable.

【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：R496;TN911.7

【相似文献】