Modeling for Control [About This Issue]

This issue of IEEE Control Systems Magazine issue features ???Signature-Based Time-Series Analysis for System Identification??? by Kourosh Danai, which discusses system identification and modeling of nonlinear dynamic systems using time-series analysis. While the use of wavelet transforms is a well-established technique for solving this problem, there are inherent drawbacks associated with the redundancy of information they generate by expansion of time series into surfaces. This article describes a new approach wherein the enhanced delineation offered by wavelet transforms is used to isolate regions of dominance, called signatures,among the wavelet coefficients of a time series.