Iterative design of uncertainty models and robust controllers based on experiment data

An iterative uncertainty modelling and μ-synthesis method is presented for LTI systems. In many robust control problems with complex dynamical systems, there is no a priori information for designing a nonconservative and viable representation of the uncertainty. The goal of this paper is to provide a practical numerical method for improving the robust performance of the controlled system by designing an appropriate uncertainty model and a controller. In the proposed iterative algorithm the controller is designed by mu-synthesis and the frequency-dependent norm-bounds of the uncertainties are tuned to minimize the structured singular value with model validation constraints. Since the uncertainty model depends on the quality of the input-output measurements, therefore closed-loop experiments are involved in the iteration in order to improve the set of validation constraints. The efficiency of the method is justified by the simulation example of a vehicle steering problem.