IDENTIFICATION FOR ROBUST MULTIVARIABLE CONTROL - THE DESIGN OF EXPERIMENTS

An approach to the design of experiments for the identification of lin ear MIMO models that will provide robust model-based control is presen ted. The design criteria, based on conditions for robust stability and performance, are developed from singular Value decompositions (SVD) c haracterization of structured model mismatch. The resulting designs mi nimize uncertainties in certain structural parameters in the SVD of th e multivariable model rather than simply the magnitude of the model mi smatch. An important feature of these designs is that input perturbati ons applied in the low-gain directions of the multivariable process ha ve much larger magnitudes than those applied in the high-gain directio ns. This leads to input sequences which are neither binary nor indepen dent. The results are used to provide an intuitive justification for p erforming identification under closed-loop multivariable control. Exam ples are presented which illustrate both the physical interpretations of the designs, and the effectiveness of using a sequential design app roach.