摘要
经验模式分解(EMD)存在的端点效应问题影响着该方法的应用。本文研究了基于端点优化对称延拓和镜像延拓的抑制EMD端点效应的改进方法,避免了单独采用端点优化对称延拓法在预测的点数较多时会造成速度较慢,以及单独采用镜像延拓法在处理端点不是极值点的短时间序列时效果不佳的问题。首先利用端点优化对称延拓法对数据序列两端各延拓一个局部极值,获取最佳的信号端点值,然后利用镜像延拓法把镜内的信号映射成一个不存在端点的环形信号,再进行经验模式分解。通过对仿真信号分析,表明该方法能够有效抑制端点效应问题。
Empirical Mode Decomposition （EMD） has a boundary effect problem which affects its application. In this paper a method to depress the boundary effect was developed based on Boundary Optimizing Symmetrical Extension method and Mirror Extension method. The developed method can resolve the issue of lower calculating speed when using Boundary Optimizing Symmetrical Extension method separately and the issue of using Mirror Extension method when the boundary of a short time series is not the extreme point. The developed method use Boundary Optimizing Symmetrical Extension method to extend a local extremum point at the end of a data series, so a best signal endpoint is got. Then the developed method use Mirror Extension method to mapping the mirror signal into a ring signal which has no endpoints. At the end the processed signal was analyzed by EMD. The result of simulation analysis shows that the developed EMD method can depress the boundary effect.
出处
《软件》
2012年第8期72-74,共3页
Software
关键词
经验模式分解
端点效应
端点优化对称延拓
镜像延拓
Empirical Mode Decomposition
Boundary effect
Boundary Optimizing Symmetrical Extension
Mirror Extension
