Adaptive minimax control: rejecting both what has happened and what might happen

摘要

Presents a state feedback receding horizon controller which is the result of re-posing and re-solving an infinite horizon minimax controller design problem at consecutive, small intervals in time. This controller is compared to the minimax controller, on which the receding infinite horizon controller is based, and an H/sub /spl infin// controller. The receding infinite horizon controller allows a trade off between reacting to the current perturbed system state and anticipating future disturbance signals. The problem is solved in the presence of structured uncertainty.