Discrete LS Estimates of Correlation Function of Bi-Periodically Correlated Random Signals
Игорь Николаевич Яворский, Oksana Dzeryn, Роман Михайлович Юзефович · Radioelectronics and Communications Systems · 2020
Abstract Analysis of discrete estimates of bi-periodically correlated random processes (BPCRP)—mathematical models of signals with double stochastic periodicity was performed using least squares (LS) technique. It is shown that LS utilization allows to avoid systematic errors related to leakage effect. Expressions for estimate bias and dispersion were obtained, allowing determination systematic and mean square errors depending on discretization step, sample number and signal parameters were obtained. For quadrature BPCRP model discrete and continuous dispersions of LS estimation of correlation components were compared. Recommendations for discretization step are given.