中文摘要：

用希尔伯特-黄变换算法（HHT）对高频周期抖动的分解进行了研究。通过将小波阈值和FFT滤波相结合,去除随机成分和低频周期抖动后,对高频周期抖动进行经验模态分解（EMD）,将得到的各固有模态函数（IMF）分量经过HHT变换得到Hilbert谱。结合IMF时域图和HHT时频谱,能较准确地估算各抖动成分的频率和其他信号特性,弥补已有算法在研究高频周期抖动的不足,并首次提出分频段分解抖动的方法。将实际抖动数据的测量结果与本方法估算的结果比较可知,HHT算法分解抖动的精度较高。

英文摘要：

High-frequency period jitter（HPJ） decomposition based on HHT algorithm is studied in this paper.Through removing the random jitter（RJ） and low-frequency period jitter（LPJ） by wavelet threshold-value and FFT filter,HPJ is decomposed by EMD,and each resulted IMF component is transformed by HHT,and thus the Hilbert spectrum is obtained.In combination of IMF time-domain diagram and Hilbert spectrum,the frequency of jitter and the other signal characteristics could be fairly estimated.This schem,for the first time,proposes the sub-band method for decomposing jitter.From comparison of the measured result and estimated result,HHT algorithm is of high precision in jitter decomposition.