Identification of continuous-time MIMO state space models from sampled data, in the presence of process and measurement noise

This paper demonstrates the use of a continuous-time subspace model identification method, in the identification of MIMO state-space models. The measured input and output signals are assumed to be measured at regularly spaced sampling instances. The presence of both measurement and process noise is considered. The proposed method gives a biased estimate of the system matrices, but we shall show how to minimise this bias by a proper choice of the lag used for the instruments. Finally, the applicability of the method is demonstrated in the identification of aircraft dynamics