Abstract
This paper proposes a procedure for recovering asymptotically the robustness of the optimal output feedback system via perfect regulation and perfect observation. The duality between the perfect regulation and the perfect observation is fully exploited to treat both of the return difference matrices at the plant input and at the plant output. Explicit expressions of these return difference matrices are derived and their asymptotic properties are examined. A simple algorithm for perfect regulation is given based on the coprime matrix fraction description. Some numerical results are given to illustrate the theory.
