Investigation and Performance Enhancement of the Empirical Mode Decomposition Method Based on a Heuristic Search Optimization Approach

Abstract

Empirical mode decomposition (EMD) is a relatively new, data-driven adaptive technique for analyzing multicomponent signals. Although it has many interesting features and often exhibits an ability to decompose nonlinear and nonstationary signals, it lacks a strong theoretical basis which would allow a performance analysis and hence the enhancement and optimization of the method in a systematic way. In this paper, the optimization of EMD is attempted in an alternative manner. Using specially defined multicomponent signals, the optimum outputs can be known in advance and used in the optimization of the EMD-free parameters within a genetic algorithm framework. The contributions of this paper are two-fold. First, the optimization of both the interpolation points and the piecewise interpolating polynomials for the formation of the upper and lower envelopes of the signal reveal important characteristics of the method which where previously hidden. Second, basic directions for the estimates of the optimized parameters are developed, leading to significant performance improvements.