Hankel Matrices and Realizations

Abstract

This paper studies the realization and the identification problems of linear time-invariant lumped systems. It is assumed that the systems to be identified can be reset to the zero states, and that the measured data are free of noise. The introduced method is based on Hankel matrices. The properties of Hankel matrices are first introduced. A new realization algorithm, which is believed to be the simplest possible, is presented. This result is then used in the identification. Both discrete-time and continuous-time systems are studied.