多周期信号的小波包方差分析方法

Wavelet packet variance analysis methods for multi-period signal

在多周期结构分析中,最大重叠离散小波变换得到的信号周期具有明显的局限性。在对比小波方差分析中,提出了用最大重叠离散小波包方差法分析不同尺度小波方差图、功率谱,从而得到信号周期估计的最大值。实验结果表明,对信号或时间序列周期结构的分析是一种有效的方法,该方法可以准确估计多周期信号的小波包方差。

In the multi - period structure analysis, the period of signal obtained by the maximal overlap discrete wavelet variance has significant limitations. In this paper, in contrast to the wavelet variance analysis, we present a method of maximal overlap discrete wavelet packet variance to analyze the variance diagram or power spectrum at different scales, and then obtain the period of signal by estimating the maximum value. Experimental results show that the method of wavelet packet variance can estimate the multi-period structure of the signal accurately. It is an ef- fective method for signal or time series periodic structure analysis.