基于改进方法EEMD的HHT脉搏信号分析

Pulse Signal Analysis Based on Improved Ensemble Empirical Decomposition of Hilbert-huang Transform

针对传统希尔伯特-黄变换(Hilbert-Huang Transform,HHT)中经验模态分解(EmpiricalMode Decomposition,EMD)存在的模态混叠问题,提出一种基于总体经验模态分解(Ensemble EmpiricalMode Function,EEMD)的脉搏信号分析方法。该方法通过对原始数据加入随机白噪声分量,使不同区域脉搏信号保持完整性,克服了传统EMD分解不能有效解决模态混叠(Mode Mixing,MM)的问题。首先通过EEMD方法提取脉搏信号的固有模态函数(Intrinsic Mode Function,IMF),再进行Hilbert变换,得到脉搏信号的Hilbert谱和边际谱。其结果可以定量并准确地刻画任意时刻的瞬时频率和幅值,为脉搏信号的特征提取和模式识别提供可靠的依据。

An Ensemble Empirical Mode Function（EEMD） method is used to analyze the pulse signal due to the Mode Mixing（MM） problem of the Empirical Mode Decomposition（EMD） in the traditional Hilbert-Huang Transform（HHT）.The EEMD method can keep the integrity of the different area pulse signal by adding random while noise into the original signal.Firstly,the EEMD is used to extract the Intrinsic Mode Function（IMF） from the pulse signal.Then,the Hilbert spectrum and the marginal spectrum of the pulse signal can be achieved by Hilbert transformation and the integration of the Hilbert spectrum,respectively.Thus,quantitative and accurate instantaneous frequency and amplitude can be achieved.It provides reliable basis for the feature extraction and pattern recognition of the pulse signal.