Empirical Mode Decomposition as a Tree-Structured Filter: A Tutorial View

摘要

Empirical mode decomposition (EMD) can adaptively decompose multicomponent signals into a finite number of simple oscillatory mode functions termed intrinsic mode functions (IMFs) through a sifting process. In this paper, a tree-structured EMD filter is designed to describe the EMD algorithm. The core element of the EMD filter is called sifting filter which depicts the EMD sifting process of extracting an IMF. As a consequence, the tree-structured EMD filter can make toward a better understanding of the EMD method.