Mixed time/frequency based Robust Identification

PABLO PARRILO; MARIO SZNAIER; RICARDO S. SÁNCHEZ PEÑA

Kobe

Conferencia; Conference on Decision and Control; 1996

IEEE

In this paper we propose a new robust identification frameworkthat combines both frequency and time-domain experimentaldata. The main result of the paper shows thatthe problems of establishing consistency of the data and ofobtaining a nominal model and bounds on the identiiicationerror can be recast as a constrained finite-dimensionalconvex optimization problem that can be efficiently solvedusing Linear Matrix Inequalities techniques. This approach,based upon a generalized interpolation theory, contains asspecial cases the Carath6odory-Fejtr (purely time-domain)and Nevanlinna-Pick (purely frequency-domain) problems.The proposed procedure interpolates the fiequency andtime domain experimental data while restricting the identifiedsystem to be in an U priori given class of models,resulting in a nominal model consistent with both sourcesof data. Thus, it is convergent and optimal up to a factorof 2 (with respect to central algorithms).