online read us now
Paper details
Number 2 - June 1999
Volume 9 - 1999
Multisine approximation of multivariate orthogonal random processes
Jarosław Figwer
Abstract
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.
Keywords
simulation random processes, multivariate orthogonal random processes, simulated identification, multisine random time-series, fast Fourier transform