Modeling of quasi-stationary signals using temporal fuzzy sets and time-frequency distributions

We present a novel method for modeling quasi-stationary signals based on time-frequency feature extraction and fuzzy classification. A family of fuzzy sets, induced on a temporal universe, is utilized to decompose a dynamic trajectory of the observed signal into patterns of physical activity. Cluster prototypes are shown to approximate the power spectra of the stationary generators that jointly govern the process realization. A simulation example is provided to illustrate the method.