基于脑电EEG的改进EEMD算法 (Improved EEMD Algorithm Based on EEG Signal)

Liya Huang, Chenglu Da, Chen Yang, Zhiyang Chen, Hao Wang

Abstract:

In order to effectively improve the mode mixing problem for EEG study,a modified EEMD algorithm was proposed to adapt to the brain signal research.Firstly,We screened out the EMD results based on correlation,then adaptively predicted the EEG signal characteristics from the original brain signals,and fused the property of white Gaussian noise to generate new noise brain signal.Finally,based on the new noise,the EEMD was performed.The experimental results show that the new brain signal is not only adaptive,but also can solve the mode mixing problem of brain signals in EMD better,proving the theory and the application value of the improved algorithm in EEG study field.

为了有效地改善模态混叠问题以适应脑电信号的研究,提出了一种改进的集合经验模态分解算法。首先对脑信号进行相关性筛选；然后自适应地从原始脑信号中预测脑电特性信号,融合高斯白噪声生成新型脑信号噪声；最后基于该噪声进行集合经验模态分解。仿真实验表明,新型脑信号噪声不仅具有自适应特性,而且可以更好地解决脑信号经验模态分解中的模态混叠问题,同时也证明了该算法在脑电研究领域的理论和应用价值。