基于屏蔽技术的经验模式分解改进方法

经验模态分解(EMD)具有优越的自适应性,对非平稳、非线性信号的不同时间尺度可进行正确的局部化处理,然而,当信号组合分量的频率相近时会出现模态混叠现象。针对该问题,提出引入信号屏蔽技术对其改进,通过向原始信号添加一定频率的屏蔽信号,将低频成分屏蔽在高频成分所在的本征模态函数(IMF)之外,而屏蔽信号自身通过加减二次分解取平均值消除对IMF分量的影响。仿真及实际应用验证了改进方法可有效克服模态混叠现象,为EMD改进提供一种新的途径。

An Improved Empirical Mode Decomposition Method Based on Shielding Technology

Empirical mode decomposition(EMD) has superior adaptability as to localize the non-stationary and non-linear signals at different time scales correctly.But there is modal aliasing phenomenon when the frequencies of signal consisting components are similar to each other.In order to solve the problem above, shielding technology is introduced to optimize traditional EMD.A shielding signal with certain frequency is added to the original signal to exclude low frequency components from in the intrinsic mode function(IMF) where high-frequency component existed, while the shielding signal itself eliminated the influence on the IMF by acquiring average after adding or subtracting decomposition twice.Simulation and practical application have proved that this improved method can overcome aliasing modes phenomenon effectively, providing a new way to improve the EMD.