Identification of Unknown Parameters of Partially Observed Discrete-time Stochastic Systems by Using a New Class of Pseudomeasurements

Abstract

Linear discrete-time stochastic systems with partially unknown system matrices are considered. Assuming that the stochastic systems are partially observable, we present a new class of pseudomeasurements modified from the one given in the preceeding paper which are fictitious and additional observation processes on the unknown entries of the system matrices. Then we propose a new class of iterative methods which identify those unknown entries and the state vectors simultaneously. A numerical example is given to demonstrate the advantages of the proposed approach.