A Practical Study of a Blind Spectrum Sensing Technique in Real Cognitive Radio Scenarios

In a cognitive radio system, spectrum sensing is considered as the key step of the cognition cycle and the primary function of the opportunistic usage of the spectrum. In real world scenarios, spectrum sensing algorithms are sensitive to several factors especially the model uncertainty problem in low signal-to-noise ratio environments. In this paper, a blind detector based on the cyclostationarity feature is implemented and tested on the USRP/GNU Radio platform. The performance of this algorithm is measured in different scenarios at various frequencies to prove the sustainability of this blind technique and its robustness against severe channel conditions. Measurements are conducted on the 802.11a bands, the 3G network, and for USRP transmitted signals. In this study, the detector based on the symmetry property of the cyclic autocorrelation function is compared to the classical energy detector. Various experimental results indicate that the blind sensing detector outperforms the energy detector by guarantying a high detection probability under real channel conditions and low signal-to-noise ratio.