COMPUTATIONAL DESIGN TECHNIQUES FOR REDUCED-ORDER OBSERVERS IN GENERALIZED STATE-SPACE SYSTEMS

In this paper we present computationally efficient design techniques f or the reduced-order observer problem in generalized state-space syste ms. The proposed approach utilizes the generalized block Hessenberg fo rm. The problem of reduced-order observer in generalized systems is re formulated as a full-order observer problem for state-space systems. T his new approach parametrizes the family of all possible solutions, st udies the problem of robustness, and relaxes the condition of observab ility.