Robust Identification of PWL-Wiener Models: Use in Model Predictive Control

Abstract

In this paper a robust identification strategy for Wiener like models constituted by a linear dynamic block a in series with a Piecewise Linear (PWL) functions as the nonlinear static gain is presented. The proposed realization allows straightforward characterization of the static gain uncertainty. A robust Model Predictive (MPC) algorithm, using the presented model strategy, is developed to guarantee that no constrains in the feedback loop are violated.