基于小波周期图的功率谱估计算法研究

Research of Power Spectrum Estimation Algorithm based on Wavelet Periodogram

作为一种功率谱的估计方法,周期图法具有物理概念清晰、直观、计算方便的优点。由于傅里叶变换的一些局限性,一直以来,经典周期图估计法不能很好的处理非平稳信号,并且缺乏很好地处理分辨率与方差的方法。文章在小波理论的基础上,结合经典周期图,深入阐述了小波周期图的基本算法与相关理论。通过matlab仿真分析表明,小波周期图有效地在各个不同的尺度上对信号的功率谱进行局部估计。

As a power spectrum estimation method,periodogram has such physical benefits as clear concept,intuition,easy calculation.Because some of the limitations of Fourier transform,the classic periodogram estimation has been incapable of handling non-stationary signals,and lack of a good deal between resolution and variance.This paper,based on wavelet theory,combined with classical periodogram,deeply elaborated the basic algorithm and related theory of wavelet periodogram.Matlab simulation results show that the wavelet periodogram can effectively give local power spectrum estimates in different scales for signals.