论文部分内容阅读
在左右手运动想象的脑电(EEG)分析方法中,目前大多针对多通道采集的EEG数据,难以应用到单通道脑机接口(BCI)中。本文采用一种改进的独立成分分析(ICA)方法,实现了对EEG数据有效的预处理。首先,对数据进行线性漂移校正,去除数据漂移;然后采用延时窗口数据增加虚拟通道数,利用ICA除去EEG数据中的伪迹,即眼电和心电;最后利用希尔伯特-黄变换(HHT)后的瞬时幅值,求平均瞬时能量特征并分类。实验证明,该方法测试性完成了EEG数据的预处理工作,提高了单通道EEG信号的分类率,可为单通道的便携式BCI研究打下基础。
In the EEG analysis method of left and right hand imagination, most of the current EEG data collected for multi-channel are hard to be applied to single-channel brain-computer interface (BCI). In this paper, we use an improved independent component analysis (ICA) method to achieve effective pretreatment of EEG data. First, the data is linearly drift corrected to remove the data drift; then the delay window data is used to increase the number of virtual channels, ICA is used to remove the artifacts in the EEG data, that is, the EEG and ECG; finally, the Hilbert-Huang transform (HHT) instantaneous amplitude, the average instantaneous energy characteristics and classification. Experiments show that this method can test the preprocessing of EEG data and improve the classification rate of single channel EEG signal, which can lay a foundation for single channel portable BCI research.