Model Set Identification Based on Euclidean Norm

Abstract The problem considered in this paper is to obtain the smallest model set which is consistent with input-output data and characterized by a nominal model, a parametric uncertainty bound and equation errors. Identification methods are proposed in two cases where (1) only structured uncertainty is considered, and (2) both structured and unstructured uncertainties are considered. The uncertainty bounds are all measured with the Euclidean norm. It is shown that the identification problems can be reduced to convex optimization problems.