OBSERVER DESIGN FOR LARGE-SCALE LINEAR SYSTEMS.

An approach is presented which reduces the computational requirements in observer design. Specifically, a procedure is developed which reduces the observer design to an algebraic problem of solving an (n minus 2m) multiplied by (n minus m) matrix equation, where n is the dimension of the state and m is the dimension of the output. It is also shown that for a special class of problems which appears very often in time-series modeling, the computational requirements can be much further reduced. The procedure developed in this paper is applicable to both discrete-time and continuous-time linear dynamic systems. The resulting observer will have m eigenvalues clustered together at a selected point and the remaining n minus 2m eigenvalues are arbitraily placed.