IDENTIFICATION OF THE DETERMINISTIC PART OF MIMO STATE-SPACE MODELS GIVEN IN INNOVATIONS FORM FROM INPUT-OUTPUT DATA

In this paper we describe two algorithms to identify a linear, time-in variant, finite dimensional state space model from input-output data. The system to be identified is assumed to be excited by a measurable i nput and an unknown process noise and the measurements are disturbed b y unknown measurement noise. Both noise sequences are discrete zero-me an white noise. The first algorithm gives consistent estimates only fo r the case where the input also is zero-mean white noise, while the sa me result is obtained with the second algorithm without this constrain t. For the special case where the input signal is discrete zero-mean w hite noise, it is explicitly shown that this second algorithm is a spe cial case of the recently developed Multivariable Output-Error State S pace (MOESP) class of algorithms based on instrumental variables. The usefulness of the presented schemes is highlighted in a realistic simu lation study.