A parametric extension of Mixed time/frequency Robust Identification

PABLO PARRILO; RICARDO S. SÁNCHEZ PEÑA; MARIO SZNAIER

Albuquerque

Conferencia; American Control Conference; 1997

American Automatic Control Council

A parametric extension to the timelfrequency robustidentification framework is presented. The result can beapplied to plants which have a clear parametric and nonparametric separation, on which time and/or frequencyexperiments have been performed. The parametric portionof the model should be &ne in the unknown parameters,which includes practical applications such as flexiblestructures. The consistency problem is cast as a constrainedfinite-dimensional convex optimization problemthat can be formulated as a Linear Matrix Inequality. Theproposed procedure provides a convergent, interpolatoryidentification algorithm.