Linear filtration methods for statistical analysis of periodically correlated random processes—Part II: Harmonic series representation

The estimates of probability characteristics for periodically correlated signals that are based on harmonic series representation are analyzed. Two methods for stationary components estimations are discussed: Hilbert transform-based method and a frequency shift method. Application of frequency shift method to real and simulated data is shown.

► New approach for cyclostationary processes probability characteristics estimation is developed.
► It is based on Hilbert transform and frequency shift method.
► Harmonic series representations with finite and infinite spectra are equivalent in the meaning of cyclostationary processes probability characteristics.
► Cross-covariances of stationary components are indicators of nonstationarity.