基于最小平方距离相关的EMD改进算法及应用

Improved method for empirical mode decomposition based on SSDA and its application

包络线的拟合是经验模式分解中的关键一步,但由于信号端点处极值的不确定性导致在信号上下包络线拟合过程中存在着严重的端点效应问题;在分析现有解决端点效应问题方法的基础上提出了基于最小平方距离相关的数据端点处包络线拟合方法。该方法在信号内部寻找与端点处信号变化趋势相关度最高的一段波形,并用此段波形的包络线来重新刻画端点处的包络线。仿真和实际信号的分析证明,基于最小平方距离相关的数据端点处包络拟合方法能够有效地抑制端点效应,算法简单并且易于实现。

Envelope fitting is the most important step in empirical mode decomposition,however,the end swing effect exists in the course of getting envelops of data because of the uncertain extremes near the ends of data.On the basis of the existing algorithms,an improved method for empirical mode decomposition based on SSDA（Sequential Similarity Detection Algorithm,SSDA） was proposed.Method finds the best correlative sequence in the inner part of data and adopts the optimum sequence to proofread the segment of envelope near the data ends.The simulation and experiment results show the method is effective in restraining end effect,furthermore the method is simple and easy to realize.