基于便携式脑电信号采集器的脑-机器人交互系统

发布时间：2018-04-01 05:39

本文选题：便携式　切入点：脑电信号采集器　出处：《电子测量与仪器学报》2016年05期

【摘要】：针对现有高精度脑电信号采集设备体积笨重、成本高、无法普及的问题,开发了便携低成本的多通道脑电信号采集器,并以可编程可定制的乐高机器人为应用实例,搭建了集注意力训练、教育、娱乐于一体的脑-机器人交互系统。脑电信号采集器内置锂电池,体积为65 mm×40 mm×21 mm,重约60 g,以无线方式传输信号。在此基础上,结合典型相关分析(canonical correlation analysis,CCA)与快速傅里叶变换(fast Fourier transformation,FFT),提出了基于CCA-Wx-FFT的SSVEP特征提取算法,对采集器信号进行分析,以确保对机器人的精确控制。10名被试的测试结果表明,本系统获得了与高精度EEG设备相近的信号特征和控制准确率(92.6%vs 94.1%)。这一脑-机器人交互系统极大地降低了系统开发的成本,方便携带,易于扩展,对脑控的普及具有重要的应用价值。
[Abstract]:In order to solve the problems of bulky, high cost and low cost multichannel EEG acquisition equipment with high precision, a portable and low cost multichannel EEG collector is developed, and a programmable and customizable Lego robot is used as an application example. A brain-robot interactive system with attention training, education and entertainment is set up. The EEG collector is equipped with a lithium battery with a volume of 65 mm 脳 40 mm 脳 21 mm and a weight of about 60 g. On this basis, the signal is transmitted wirelessly. Combined with canonical correlation analysis and fast Fourier transform, a SSVEP feature extraction algorithm based on CCA-Wx-FFT is proposed to analyze the signal of the collector to ensure the accurate control of the robot. In this system, the signal features and control accuracy are similar to those of high-precision EEG devices. This brain-robot interaction system can greatly reduce the cost of system development, be easy to carry and expand. It has important application value to the popularization of brain control.
【作者单位】：天津大学电气与自动化工程学院;三江大学工程教育学院;加州州立大学贝克斯菲尔德分校计算机科学与电子工程系;
【基金】：国家自然科学基金(61473207)资助项目
【分类号】：TP11

【参考文献】