两类运动想象脑电识别中小波变换对信源适用状况的研究

he paper studied the wavelet transform feature extraction of EEG signal. To exert its effects of data analysis, the technique of wavelet transform needs suitable conditions. Among them, the modes of EEG signal source is the important items to be considered. Based on the two types of motor imagery EEG data of the international BCI competitions, the research selected the single channel EEG signals, two-channel EEG signals, EEG signals power etc as data source, also selected the wavelet transform with db basis function as the analysis tool, to verify the signal analysis ability of the wavelet transform technique. The research found that, although increased channel amount can have the opportunity to improve the efficiency of EEG identification, not all kinds of multi-channel signals were applicable to wavelet transform feature extraction, therefore we need to consider how to use multi-channel signal source reasonably. Moreover, although the signal power can be the important data source to EEG identification, it seemed that such data source cannot get on well with the process of wavelet transform, hence, it led to the reduction of identification efficiency. To all kinds of EEG signal modes, we need to make detailed analysis, also study various factors which influence the feature extraction of wavelet transform, then use wavelet transform in proper way in order to give full play to the advantages of this technique in suitable conditions.