The conditionality of Gram’s information matrix in the task of identification: the scaling of input and output signals

Download article in PDF format

Authors: Abdenova G. A., Voevoda A. A.

Annotation: The scaling procedure of the input and output data for raising the least-squares estimators’ accuracy in the tasks of the parametric identification of the objects which have been by the models in the space state is considered. The proposed procedure of the scaling allows eliminating the property influence of the small changes of the filtration estimations in reference to some constant. Results of calculations of conditionality numbers of information matrixes Grama received by means of estimations of a filtration of component of vector of a condition without scaling and according to estimations with preliminary procedure of scaling are shown.

Keywords: model in the space state, parameters identification, gram’s matrix, the conditionality number, scaling