ITERATIVE LQG CONTROLLER-DESIGN THROUGH CLOSED-LOOP IDENTIFICATION

This paper presents an iterative LQG controller design approach for a linear stochastic system with an uncertain open-loop model and unknown noise statistics. This approach consists of closed-loop identificatio n and controller redesign cycles. In each cycle, the closed-loop ident ification method is used to identify an open-loop model and a steady-s tate Kalman filter gain from closed-loop input/output test data obtain ed by using a feedback LQG controller designed from the previous cycle . Then the identified open-loop model is used to redesign the state fe edback. The state feedback and the identified Kalman filter gain are u sed to form an updated LQG controller for the next cycle. This iterati ve process continues until the updated controller converges. The propo sed controller design is demonstrated by numerical simulations and exp eriments on a highly unstable large-gap magnetic suspension system.