文章摘要

文章信息/Info

Title:: Classification of EEG motor imagination based on wavelet packet decomposition and common space model

Author(s):: WANG Mei-e1; XU Yanhua2; 1.Henan Provincial Hospital，Zhengzhou 450000，China；2. Zhongyuan University of Science and Technology ，Zhengzhou 450000

Keywords:: Wavelet packet decomposition; The common space model method; Brain electrical motor imagination; De noise; Decision tree; Classification method

摘要:: 针对脑电运动想象信号分类不准确，本研究结合小波包分解和共空间模式，提出了一种脑电运动想象分类方法。采用电极无创式采集脑电运动信号，并利用独立成分分析去噪，通过小波变换和共空间模式，提取脑电运动信号的时域特征和空间域特征。以决策树算法为基础，一个决策树对应一种脑电运动信号类别，构建随机森林分类器，实现脑电运动想象分类。结果表明：与其他分类方法相比，该方法能够较好地控制智能小车的运动反应，证明本研究方法的分类质量更高。

Abstract:: Aim to the inaccurate classification of EEG motor imagery signals, we proposed a classification method of EEG motor imagery based on wavelet packet decomposition and common spatial pattern. The EEG motor imagery was collected noninvasively by electrodes, and was denoised by independent component analysis (ICA). The temporal and spatial features of EEG motor imagery were extracted by wavelet transform and common spatial pattern. Based on the decision tree algorithm, a decision tree corresponded to a category of EEG motor signals, and a random forest classifier was constructed to realize the classification of EEG motor imagery. The results showed that compared with other classification methods, this method can better control the motion response of the intelligent car. It proves that the classification quality of the proposed method is higher.