Active parametric identification of stochastic continuous-discrete systems obtained by statistical linearization

V. I. Denisov, V. M. Chubich, E. V. Filippova

Novosibirsk State Technical University, Novosibirsk, Russia

Abstract: For stochastic continuous-discrete models obtained by statistical linearization, we consider the theoretical and applied aspects of active parametric identification in the case where the parameters of mathematical models to be estimated occur in the state, control, and perturbation matrices and the covariance matrices of dynamic noise and measurement errors. The results are given. An example of optimal parameter estimation for one model structure is considered.

Keywords: parameter estimation, maximum likelihood method, design of optimal input signals, information matrix, optimality criteria.

UDC: 519.24+519.711

Received: 13.06.2012

Bibliographic databases:

•