Performance Enhancement of Cyclostationarity Detector by Utilizing Multiple Cyclic Frequencies of OFDM Signals

Abstract

Spectrum sensing is a key technical challenge for the cognitive radio (CR) technology which allows it to access the spectrum of the licensed systems without causing interference to them. It is well known that cyclostationarity detectors have a great advantage of the robustness of noise uncertainty which significantly degrades the performance and makes its implementation difficult in energy detectors. This paper pays attention to the fact that cyclostationarity detector can acheive diversity gain by manipulating multiple cyclic autocorrelation functions (CAFs). While most of combining schemes in cooperative sensing with multiple detectors require the signal to noise power ratio estimation which is impractical in low SNR regime, combining with multiple CAFs in single detector based on transmit signal characteristic can be used. In this paper, three detector schemes of selection combining (SC), equal gain combining (EGC) and maximum ratio combining (MRC) with multiple CAFs are evaluated in additive white Gaussian noise (AWGN) channel considering the OFDM signal of Japanese digital television broadcasting (ISDB-T) as a primary system. The numerical results will show that the cyclic detector based on MRC using several cyclic frequencies has the best detection performances even though EGC shows slightly worse performance in assumed condition.