Closed Form Solution for Minimum Norm Model-Validating Uncertainty

A methodology in which structured uncertainty models are directly constructed from measurement data for use in robust control design of multivariable systems is proposed. The formulation allows a general linear fractional transformation uncertainty structure connections with respect to a given nominal model. Existence conditions are given, and under mild assumptions, a closed-form expression for the smallest norm structured uncertainty that validates the model is given. The uncertainty bound computation is simple and is formulated for both open and closed loop systems.