On AR representations for cyclostationary processes

Author

Sherman, P.J. ; White, L.B. ; Bitmead, R.R.

Author_Institution

Iowa State Univ., Ames, IA, USA

Volume

4

fYear

1993

fDate

27-30 April 1993

Firstpage

260

Abstract

The authors consider autoregressive (AR) types of wide sense cyclostationary (WSC) processes. The comparative performance of the discrete Fourier transform (DFT) autoregressive (AR) methods of estimating the time-periodic spectral density of an AR(2) WSC process is provided. Problems with these methods are addressed in the case of uncertainty of the process period. Examples concerning an AR(2) process subjected to period drift and randomness are provided to show that the time-varying spectral estimate converges to a time-invariant one. Results from stochastic differential equations which support this behaviour are cited. Finally, the method of extended Kalman filtering is proposed to track a slowly time-varying period.<>

Keywords

Kalman filters; differential equations; fast Fourier transforms; random processes; signal processing; tracking; autoregressive representations; cyclostationary processes; discrete Fourier transform; extended Kalman filtering; period drift; randomness; stochastic differential equations; time-periodic spectral density; uncertainty;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1993. ICASSP-93., 1993 IEEE International Conference on

Conference_Location

Minneapolis, MN, USA

ISSN

1520-6149

Print_ISBN

0-7803-7402-9

Type

conf

DOI

10.1109/ICASSP.1993.319644

Filename

319644