A Simple Method to Enhance the Detection of Second Order Cyclostationarity

Cyclostationarity is a significant feature of numerous physical and man-made processes. Specially in wireless communications, it has been explored and applied for many processing, like timing and frequency synchronization, channel estimation, and modulation classification. In this paper, we notice that the second order time varying auto-correlation of the received communication signal contains a cyclostationary real and a non-cyclostationary complex part. We propose a simple method to combine the positive and negative part of the cyclo-spectrum (the Fourier transform of the time varying autocorrelation) to enhance the performance of detecting the second order cyclostationarity of these signals. The idea redis related to the concept of cyclic Wiener filter put forth by Gardner in 1993 [4]. Monte Carlo simulations are given to verify the results.