Modeling of two-dimensional systems using cumulants

Author

Valenzuela, Hector M. ; Salvia, A. David

Author_Institution

Dept. of Electr. & Comput. Eng., Pennsylvania State Univ., University Park, PA, USA

fYear

1991

fDate

14-17 Apr 1991

Firstpage

2913

Abstract

The problem of obtaining the ARMA parameters of a 2-D system using slices of the output cumulant is considered. It is shown that cumulant slices can be used to identify the AR part of the model only if the system is causal and the denominator is product separable. With these restrictions, a computationally feasible method of determining the order of the denominator polynomials and finding the denominator polynomial coefficients is obtained. Existing methods can then be used to produce the numerator polynomial coefficients. The method is then verified through simulations

Keywords

multidimensional systems; parameter estimation; polynomials; signal processing; statistical analysis; ARMA parameters; cumulant slices; denominator polynomial coefficients; modelling; numerator polynomial coefficients; output cumulant; parameter estimates; signal processing; system identification; two-dimensional systems; Additive noise; Autocorrelation; Computational complexity; Computational modeling; Gaussian noise; Higher order statistics; Polynomials; Signal processing; Signal processing algorithms; System identification;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1991. ICASSP-91., 1991 International Conference on

Conference_Location

Toronto, Ont.

ISSN

1520-6149

Print_ISBN

0-7803-0003-3

Type

conf

DOI

10.1109/ICASSP.1991.151012

Filename

151012