面向人机交互的运动想象脑电信号感知算法研究

脑电信号的特征提取与分类识别是脑机交互领域的核心问题。针对运动想象脑电信号的多分类问题,以更好利用包含有用信息的脑电信号频带为目的,提出了基于小波包变换(WPD)和一对多共空间模式(CSP)的特征提取算法。首先使用WPD算法将原始脑电信号分解成一系列子频带,筛选与运动想象活动相关的子频带。然后使用一对多CSP算法进行特征提取。最后对各子频带的特征进行组合并使用BP神经网络进行分类。算法的有效性通过BCI竞赛的基准数据集进行了测试,相交于竞赛结果有了明显提升。

Research on Perceptual Algorithm of Motor Imagery EEG for Human-Computer Interaction

The feature extraction and classification of EEG signals are the core issues in the field of brain computer interaction.To make full use of the EEG frequency bands containing useful information,a feature extraction algorithm based on wavelet packet transform(WPD)and one vs rest common space pattern(CSP)is proposed.Firstly,WPD algorithm is used to decompose the original EEG signal into a series of sub-bands,and based on this the sub-bands related to the activity of motion imagination are screened out.Then one vs rest CSP algorithm is used to extract EEG features.Finally,the characteristics of each sub-band are combined and classified by BP neural network.The validity of the algorithm is tested by the benchmark data set of BCI competition,and it is shown that the result is obviously improved in comparison with the competition result.