Digital implementation analysis of cyclostationary algorithms in cognitive radio

Spectral awareness is currently one of the most challenging problems in CR designer. Recent researches in this area have suggested that cyclostationary approach is proposed as an effective method to solve the detection and estimation in spectral awareness. The drawback of this theory is its complexity of digital implementation. This paper specializes four realizations, namely, SSCA algorithm, FAM algorithm, direct estimation of cyclic periodogram, SCD method. The main issues involve in their detection effects (which are described by detection probability of false alarm), computational efficiencies for spectral correlation estimation and their high-speed digital realizations. The theoretical framework is characterized as a periodically time-varying quadratic system with a kernel function. Conclusions of comparison demonstrate both SSCA algorithm and direct estimation of cyclic periodogram account to low signal-to-noise (SNR=-10dB) scenarios. SSCA algorithm is fit for fast detection, while direct estimation is suitable for parameter estimation.