Krasnik A. The adaptive dynamic multivariate digital filter use for accuracy prediction improving in management systems

Українська версія

Thesis for the degree of Candidate of Sciences (CSc)

State registration number

0408U003573

Applicant for

Specialization

  • 05.13.06 - Інформаційні технології

11-07-2008

Specialized Academic Board

Д.26.001.40

Essay

The thesis is dedicated to problems of increase of accuracy of forecasting on the basis of usage of digital adaptive dynamic multidimensional filters (DADMF). The frame linear DADMF is offered on the basis of approximating obscure weighting function of object, the parameters which one are forecast, polynomials (Chebyshev, Ermita, Laguerre etc.). For definition of unknowns of factors of approximating the method of least squares utilised. The theories of operation of offered frame of the filter for multidimensional objects and collimating to the filter of adaptive properties are designed. The influencing on accuracy of number of supervision, number of approximating polynomials and exact guessing of an interval of a lag effect is investigated. The quantitative assessments allow to reduce time to synthesis of the filter and to increase accuracy of forecasting of behavior of parameters of inertial non-steady multidimensional objects.

Files

Similar theses