Korbicz, Józef - red. ; Uciński, Dariusz - red.
Multisine approximation of multivariate orthogonal random processes
Group publication title:
Subject and Keywords:
simulation random processes ; multivariate orthogonal random processes ; simulated identification ; multisine random time-series ; fast Fourier transform
An approach to the synthesis and simulation of wide-sense stationary multivariate orthogonal random processes defined by their power spectral density matrices is presented. The approach is based on approximating the non-parametric power spectral density representation by the periodogram matrix of a multivariate orthogonal multisine random time-series. ; This periodogram matrix is used to construct the corresponding spectrum of the multivariate orthogonal multisine random time-series (synthesis). Application of the inverse finite discrete Fourier transform to this spectrum results in a multivariate orthogonal multisine random time-series with the predefined periodogram matrix (simulation). The properties of multivariate orthogonal multisine random process approximations obtained in this way are discussed. Attention is paid to asymptotic gaussianess.
Zielona Góra: Uniwersytet Zielonogórski
AMCS, volume 9, number 2 (1999) ; click here to follow the link