Minimum variance constrained estimator

This paper is concerned with the problem of state estimation for discrete-time linear systems in presence additional (equality or inequality) constraints on (or estimate). By use minimum variance duality, converted into an optimal control problem. Two algorithmic solutions are described: full information estimator (FIE) and moving horizon (MHE). The main result to show that proposed stable sense observer. algorithm distinct from standard constrained based upon energy duality. two compared numerically benchmark batch reactor process model.