基于EEMD和混沌的信号特征提取方法及应用

A Signals Feature Extraction Method Based on the EEMD and Chaotic and Its Application

提出了一种基于集合经验模态分解（EEMD）和混沌相结合的信号特征提取方法，应用于婴儿呼吸信号哮喘检测中。EEMD把呼吸的局部信号分解成一系列频率从高到低的模态分量，对各分量与局部呼吸信号进行相关分析，并给出各分量的Hilbert谱，通过局部分析的结果初步判断婴儿是否患有哮喘；然后，以EEMD局部信号检测出来的信号频率作为混沌振子检测的频率，对全局呼吸信号进行整体检测及分析，由混沌的间歇周期可以得出原始呼吸信号的频率，准确确定婴儿哮喘诊断结果。对EEMD和混沌算法的应用存在的问题进行了改进，将其应用到实测信号的分析中，验证了方法的有效性。该方法能够正确地反映信息特征，准确率高。

A method of feature extraction combining ensemble empirical mode decomposition （EEMD） with chaotic, and its application to the asthma detection in infants breathing signal are described. The partial infant breathing signal was decomposed by EEMD into a series of frequency mode components, which spread from high frequency components to low ones. The correlation between each component and partial signals was analyzed and the Hilbert spectral for major components was listed. From the partial signal, it may be generally determined whether the child suffered asthma. Then, the frequency calculated by EEMD from partial signals was chosen as the chaotic oscillator frequency to detect and analyse global breathing signals. It could express the original breathing signal frequency from the intermittent cycle of chaotic timedomain waveform, and confirm the result for infant asthma detection eventually. The EEMD and chaos algorithms were improved in order to succeed in their application. The method was put into the real data analysis and its efficiency was verified. It reflects the signals information correctly and has a high accuracy.