Empirical Mode Decomposition for vectorial bi-dimensional signals

Abstract

In this work, an Empirical Mode Decomposition (EMD) of bivariate signals is proposed. It permits to decompose signals into elementary rotating functions, oscillating functions and tendencies. The technique is based on recursive extraction of non planar and rapidly rotating functions, and then a recursive extraction of rapidly oscillating functions as in case of univariate EMD. Our concept is based on a theoretical decomposition and is confirmed on simulated signals.