Identification for multirate multi-input systems using the multi-innovation identification theory

This paper considers identification problems of multirate multi-input sampled-data systems. Using the continuous-time system discretization technique with zero-order holds, the mapping relationship (state–space model) between available multirate input and output data is set up. The multi-innovation identification theory is applied to estimate the parameters of the obtained multirate models and to present a multi-innovation stochastic gradient algorithm for the multirate systems from the multirate input–output data. Furthermore, the convergence properties of the proposed algorithm are analyzed. An illustrative example is given.