Stationary and cyclostationary random process models

Abstract

Cyclostationary random process modeling is an area of signal processing that has been the subject of numerous journal papers. W.A. Gardner (1988) devoted half of a book to cyclic spectral analysis. Cyclostationarity, however, has not received very much attention at regional conferences in the recent past, which implies that it is not being utilized by many practising engineers. Therefore, this paper reviews both stationary and cyclostationary random process models. It will be seen that cyclostationary models are more complete than stationary random process models for many manmade signals. Since these signals are best modeled as cyclostationary random processes, signal processors that exploit cyclostationarity can, in principle, have performance superior to traditional processors that utilize only stationary statistical models.<>