On quantitative a priori measures of identifiability of coefficients of linear dynamic systems

Authors:
A. A. Lomov
Affiliations:
Institute of Mathematics, Siberian Division, Russian Academy of Sciences, Novosibirsk, Russia 630090
Venue:
Journal of Computer and Systems Sciences International
Year:
2011

Citing 4
Cited 1

System identification: theory for the user

System identification: theory for the user
From time series to linear system—Part III. Approximate modelling

Automatica (Journal of IFAC)
Optimization of Parameters of Stationary Models in a Unitary Space

Automation and Remote Control
Comparison of methods for estimating the parameters of linear dynamic systems by measurements of short sections of transient processes

Automation and Remote Control

Quantile based decision making rule of the neural networks committee for ill-posed approximation problems

Neurocomputing

Quantified Score

Hi-index	0.00

Visualization

Abstract

A number of quantitative a priori measures of identifiability of coefficients of linear dynamic control systems are proposed. The criteria are based on numerical characteristics of the lower bound of the asymptotic variance of estimates of the coefficients, which is approximately calculated using the mean value of the inverse information matrix. Information matrices are calculated in the limit case of small noise amplitude. Additive observation noise is considered. In this case, the optimal estimates of coefficients correspond to the smallest distance between observations of trajectories and the solution set of the system. Numerical examples are given.