经验模态分解中虚假模态分量消除法

Effect of Sampling on Empirical Mode Decomposition and Correction

基于能量原理,提出经验模态分解(EMD)中虚假模态分量消除方法。在正常采样的条件下,分析信号EMD分解误差与虚假模态分量的关系,讨论虚假模态分量的性质。从EMD分解的完备性角度考虑在虚假模态存在情况下能量不守恒,给出模态函数消除法,从EMD分解结果中剔出虚假模态分量,消除信号经验模态分解误差。其有效性在所给例子中得到证实。

A method for cancelling pseudo mode function in empirical mode decomposition(EMD) is presented based on energy conservation.The properties of pseudo mode function are investigated on the relation between EMD error and the pseudo mode function under normal sampling.Using the completeness,energy conservation and properties of EMD,the principle of mode function cancellation is proposed for ticking out the pseudo mode functions from the decomposition results and cancelling the error in the intrinsic mode function.Moreover,an example illustrates the validity of the method.The research study has theoretical value and engineering significance for EMD processing at high accurate level under the condition of unreasonable sampling.