Closed and open loop subspace system identification of the kalman filter

Abstract

Some methods for consistent closed loop subspace system identification presented in the literature are analyzed and compared to a recently published subspace algorithm for both open as well as for closed loop data, the DSR-e algorithm. Some new variants of this algorithm are presented and discussed. Simulation experiments are included in order to illustrate if the algorithms are variance efficient or not. © 2009 Norwegian Society of Automatic Control.