Interval contractor-based reference governor for a class of uncertain nonlinear systems

This paper proposes a novel interval contractor-based reference governor design for a class of nonlinear systems containing both polynomial terms and unknown constant parameters which are valued in known bounded intervals. The reference governor ensures that input and state polynomial constraints are satisfied, despite uncertainties in the model and constraints. Its synthesis computes a maximal output admissible set (or MOAS), and requires adding new components to the state vector. It also uses interval analysis methods to shrink the bounds for the uncertain parameters over time. The MOAS computations are performed offline using a grid of the uncertain parameter bounds. Then, the reference governor is computed online by combining bisection algorithms and interval contractors.