Time-frequency analysis system based on reassignment with multitapering

This paper introduces a time-frequency (TF) analysis system we developed on the platform of MATLAB. The toolset includes common tools for evaluating statistical characteristics of nonstationary signals. Recently, Xiao and Flandrin introduced a highly optimized method for spectrum estimating, which we called Xiao-Flandrin algorithm. It combines the TF reassignment with multitapering. We intergrates it into our system. We also embed several classical methods so that a user may select a specific TF analysis method he or she prefers. On the other side, those functionalities of the system can be used for the performance comparisons. The present system has a user-friendly interface, allowing the user to import or export datas through dialog boxes and to use operation buttons and pull down menus. In addition, we demonstrate the utility of this toolset by analyzing a two component chirp signal and a bending moments recording of encountered waves in four methods. The results show this system may yet be new in the sense that the recent Xiao-Flandrin TF algorithm is included for experimenting, teaching and researching.