Title :
On test horizon for model validation by output error
Author :
Duong, Hoai Nghia ; Landau, Ioan Doré
Author_Institution :
Lab. d´´Autom., CNRS, Grenoble, France
fDate :
1/1/1994 12:00:00 AM
Abstract :
We derive necessary and sufficient conditions under which the parameters of the model are identical to those of the true system which is supposed to be linear and of finite dimension. These conditions are that the cross-correlations between the output error and the input as well as the model output should be asymptotically null in a finite horizon. This horizon may be used as a minimum necessary test horizon in model validation
Keywords :
identification; linear systems; modelling; multidimensional systems; asymptotically null cross-correlations; linear finite dimensional systems; model validation; necessary and sufficient conditions; output error; test horizon; Delay systems; Instruments; Linear systems; Predictive models; Sufficient conditions; Testing; White noise;
Journal_Title :
Automatic Control, IEEE Transactions on
