Two-dimensional FFT Algorithm for Estimating Spectral Correlation Function of Cyclostationary Random Processes
Timofey Ya. Shevgunov, Evgeniy Efimov · 2019
This paper presents the DSP algorithm for the estimation of the spectral correlation function describing properties of cyclostationary random processes in the bifrequency domain. The proposed algorithm avoids any segmentations applied to the long observation of the process measured in discrete time. Instead, it is based on two-dimensional processing of the sample autocorrelation matrix - its windowing and further application of the two-dimensional fast Fourier transform. The simulation carried out with the pulse-amplitude modulated process exhibiting strong cyclostationarity has proved the efficacy of the algorithms.