Signal Cyclostationarity Detection Based Spectrum Sensing Using Simple Diversity Combining Technique and Its Implementation

This paper presents a simple equal gain combining technique for cyclostationarity detection based spectrum sensing in cognitive radio networks. The presented technique is based on maximum cyclic autocorrelation function (MCAS) technique with a low computational cost relative to other cyclostationarity detection based spectrum sensing techniques. MCAS judges whether an orthogonal frequency division multiplexing signal is included in received signals, by comparing the peak and non-peak values of a cyclic autocorrelation function (CAF). In this paper, the signal-to-noise ratio (SNR) of the CAF, which is composed of the peak and non-peak values of the CAF, is employed. The presented technique attempts to improve the performance of spectrum sensing by combining CAFs obtained at each receive antenna and by obtaining a stable CAF SNR. Moreover, the presented technique is implemented on a software defined radio based testbed for the evaluation. The developed testbed is mainly composed of a Universal Software Radio Peripheral/GNU Radio which is one of the software defined radio receiver, and the spectrum sensing technique is experimentally demonstrated in an anechoic chamber using the 470–710 MHz frequency band allocated to ISDB-T (terrestrial digital broadcasting) in Japan. The effectiveness of the presented techniques is validated by these results.