State matrix identification via Poisson moment functionals

Abstract

Poisson moment functional are used here to identify the state matrix of linear time-invariant systems. The method provides an effective system identification algorithm in the presence of arbitrary unknown initial conditions of control systems. The method has the desirable attributes of good accuracy, direct sample data, case of calculation, and suitability for state-space models and arbitrary initial conditions. Examples are included to illustrate the applicability of the proposed method.