黄变换的穿越筛分法

Crossing Sifting Method Based on Huang Transform

基于黄变换提出了一种分解非线性、非平稳时间序列的穿越筛分方法,该方法先搜索到信号的局部极值点,然后定位出相邻局部极值点间的穿越点,最后使用三次样条对穿越点列插值,可近似得到信号的包络中值。通过实例比较分析了穿越筛分法与黄变换的经验模态分解方法,筛分结果表明该方法简单有效,可以从观测时间序列中筛分出较好的各阶固有模态函数。

Crossing sifting method is developed based on Huang transform, which is used for analyzing nonlinear and non-stationary time series. The novel method can obtain the approximate mean of signal upper and lower envelopes directly by interpolating the crossing series using cubic spline function. After finding the local extrema of time series, the crossing series can be located by neighbor extrema. The empirical mode method of Huang Transform and new method are compared through several examples. The simulation results show the new method is a simple and valid one, and it can decompose a signal into some satisfying intrinsic mode functions.