An Adaptive Iterative Filtering Method for Signal Decompositions and Instantaneous Frequency analysis.

Haomin Zhou
Georgia Institute of Technology
School of Mathematics

The empirical mode decomposition (EMD) was a method pioneered by Huang et al, as an alternative technique to the traditional Fourier and wavelet techniques for nonlinear and non-stationary signals. It decomposes a signal into several components called intrinsic mode functions (IMF), which have shown to admit better behaved instantaneous frequencies via Hilbert transforms. In this talk we present our recent progress on an alternative algorithm for EMD based on iterative filtering (IF) method. The filters are generated by the Fokker-Planck equations, second order linear PDEs for diffusion processes. This method is highly nonlinear and data dependent, and it is easy to be implemented and generalized to higher dimensions. It yields similar results as the traditional sifting algorithm for EMD. We also give the conditions under which the IF method converges, and presents a different way to perform instantaneous frequency analysis. This talk is based on collaborated work with Jingfang Liu(Georgia Tech) and Antonio Cicone (Georgia Tech).

Presentation (PDF File)

Back to Adaptive Data Analysis and Sparsity