Algorithm for Estimating the Spectral Correlation Function Using the 2D Fast Fourier Transform

An algorithm for estimating the spectral correlation function of cyclostationary random processes on the basis of their finite-time realization is proposed. The main algorithm steps are described and the main operations performed at each of them are formalized. The presented estimates of the spectral correlation function in the presence of a correlated stationary noise overlapping the analyzed cyclostationary random process in the frequency domain confirm the efficiency of the algorithm with a low signal-to-noise ratio. The number of required computational operations and the amount of computer memory required for storing intermediate results are estimated.