A new cyclostationary spectrum sensing approach in cognitive radio

In this paper, we model an almost cyclostationary signal as Fourier series expansion with lowpass time-varying coefficients and then provide a new approach to detect a cyclostationary signal in spectrum sensing cognitive radio paradigm. The lowpass time-varying coefficients of the cyclostationary signals are expanded by prolate spheroidal sequences (DPSS) as the unconditional orthonormal bases which can perfectly represent all lowpass signals. This expansion is utilized to suggest a novel MMSE estimator. This paper also explores the capabilities of the new estimation procedure for almost cyclostationary based blind detector in (non)-Gaussian noise as opposed to conventional cyclostationary modeling. Simulations are presented to compare the new method with conventional cyclostationary detector in spectrum sensing cognitive radio detection.