A receding-horizon approach to robust output feedback control for nonlinear systems

Describes a receding horizon approach to solving nonlinear dynamic games related to robust control. The controller has a time-varying perturbation-feedback structure. Open-loop trajectories are calculated to solve for control inputs using output measurements, then linear feedback is applied to the perturbations of the output versus the optimal output trajectory. Conditions are given which will yield such a controller under conditions similar to those for the existence of an optimal nonlinear observer-feedback controller. The initial and terminal weighting matrices are used to guarantee stability and to bound the cost.