Abstract
We consider partially observed discrete-time linear stochastic systems and assume that some entries of the system matrices are unknown. We propose a new method which identifies those unknown entries and the state vectors of those systems simultaneously. The key idea of the proposed method is utilization of the pseudomeasurement which is a fictitious and additional observation process on the unknown entries and will be modified so as to work for the partially observed systems. Augmenting the pseudomeasurement with the original observation process, we derive the new identification method by applying the extended Kalman filter. The proposed method is consistent with the conventional method (without pseudomeasurement) and so they can easily be unified to be a single iterative process for simultaneous identification and state estimation by switching the coefficient matrix of the augmented observation process.
