A Simulation Study on the Detection of a Cyclostationary Signal buried in a Stationary Noise for Unknown Power Scenario

This paper deals with the detection problem aiming at revealing the presence of a wide-sense cyclostationary process in correlated wideband and narrowband stationary noises while the power of both processes assumed unknown. It is shown incapability of using energy detector in such scenario. The six cyclostationary detector statistics based on a single slice taken along the cyclic frequency out of the estimated spectral correlation function are under consideration. The performance of the detectors implementing the proposed statistics is evaluated and compared via Monte Carlo simulations by example of a base-band BPSK signal. The performance advantages are measured in terms of receiver operating curves and the probability of detection versus signal-to-noise ratio