本文關(guān)鍵詞：基于改進(jìn)EEMD的呼吸波提取算法　出處：《科技通報》2017年01期 　論文類型：期刊論文

  更多相關(guān)文章： 改進(jìn)EEMD 瞬時頻率 呼吸波 容積脈搏波

【摘要】：針對人體指端的光電容積脈搏波(PPG)提取呼吸波的缺點(diǎn),并根據(jù)PPG信號的特點(diǎn)提出一種改進(jìn)的整體平均經(jīng)驗?zāi)B(tài)分解(EEMD)算法。該算法利用PPG信號次波峰幅度平均值來確定添加噪聲強(qiáng)度,計算每層分解的平均瞬時頻率,在減少分解層數(shù)的同時選取較好的固有模態(tài)函數(shù)(IMF)重構(gòu)呼吸波。采用MIMIC數(shù)據(jù)庫中的4組數(shù)據(jù)進(jìn)行測試,得到波形相對相關(guān)系數(shù)在0.65以上,幅度平方相干系數(shù)在0.98以上。實驗結(jié)果表明,本文算法不但克服了EEMD算法提取呼吸波添加噪聲和循環(huán)次數(shù)的不準(zhǔn)確性,而且可以更有效、更準(zhǔn)確的提取PPG中的呼吸波。
[Abstract]:The drawback of extracting respiratory wave is the photoelectric volumetric pulse wave (PPG) of human finger. According to the characteristics of PPG signal, an improved global average empirical mode decomposition (EMD) algorithm is proposed, which uses the average of the sub-peak amplitude of PPG signal to determine the added noise intensity. The average instantaneous frequency of each decomposition layer is calculated and the respiratory wave is reconstructed by selecting a better inherent mode function (IMF) while reducing the number of decomposition layers. Four groups of data in the MIMIC database are used for testing. The relative correlation coefficient of waveform is above 0.65, and the squared coherence coefficient of amplitude is above 0.98. The experimental results show that the correlation coefficient is more than 0.65. The proposed algorithm not only overcomes the inaccuracy of the EEMD algorithm in extracting the respiratory wave adding noise and the number of cycles, but also can extract the respiratory wave from PPG more effectively and accurately.
【作者單位】： 北京信息科技大學(xué)通信工程系;
【基金】：國家科技重大專項煤層氣田地面集輸信息集成及深度開發(fā)技術(shù)(2011ZX05039-004-02) 北京信息科技大學(xué)研究生科技創(chuàng)新項目(5111524104)
【分類號】：R443.6;TN911.6
【正文快照】： 呼吸是人體生命體征的重要參數(shù)之一,也是危重病人首要的和持續(xù)的檢測參數(shù)。目前,呼吸波信號通常使用呼吸量測定法等傳統(tǒng)方法,這些方法不但會干擾患者的自然呼吸而且無法在壓力測試、睡眠研究、家庭醫(yī)療等多種情況下使用�；诠怆娙莘e脈搏波(photoplethysmograph-ic,PPG)可以，

本文編號：1408946