Comparison between intrinsic time-scale decomposition and Hilbert-Huang transform

Abstract

The paper introduces a new non-stationary and nonlinear signal processing method: intrinsic time-scale decomposition (ITD). The ITD method can self-adaptively decompose a complex signal into a set of proper rotation components (PRCs) with really physical meaning and one monotonous trend; it constructs the piece-wise linear baseline signal between successive extrema and computes the instantaneous amplitude and frequency based on the piece-wise wave of each PRC. First, the ITD method is presented. Then, it is used to analyze a simulation signal and good results are obtained. Finally, the ITD method is compared with the Hilbert-Huang transform method and the comparison shows that the ITD method has some clear advantages in preventing overshoots and undershoots, reducing the boundary effect and increasing the computation efficiency.