Reference Governor for Constrained Systems with Time-varying References

This paper proposes a reference governor construction method for constrained linear systems with time-varying references. The main advantage of this reference governor is to simultaneously consider fulfillment of constraint conditions and evaluation of tracking performance. To derive a condition for constraint fulfillment we adopt a concept of maximal output admissible sets, and to evaluate tracking performance the management problem of time-varying references is reduced into a convex quadratic programming problem. Finally, the reference governor is obtained in the form of a piecewise affine function of state and reference variables, using a multi-parametric programming technique. In addition, we demonstrate the effectiveness of reference governors by a numerical example of a practical DC position servomechanism with control constraints