Identification of Partially Unknown System Matrix From Noisy Observation Data via Pseudomeasurement Approach

Abstract

In this paper, an identification method of the state-space model is proposed. Proposed method assumes that some entries of system matrix A is unknown and identify these entries from input and output data. The key idea of the proposed method is the use of pseudomeasurement which is fictitiously constructed data as if it were made on the unknown entries. Augmenting this pseudomeasurement with original observation data, unknown entries (and state-vector) are identified by the extended Kalman filter.