On the implementation of reference governor

Abstract

This paper proposes an implementation method of reference governor. If the constrained variables at the steady state are 0, it is rather easy to compute the maximal output admissible set Ω∞. However, if the constrained variables at the steady state are not 0, then computing Ω∞ is very difficult. We do not compute the Ω∞, but we propose a parametric characterization of Ω∞. By using this characterization, we propose an algorithm to decide reference signal r[k] on-line so that constraints are not violated. The usefulness of the proposing method is examined through numerical examples.