A Spectrum Sensing Algorithm Based on Stochastic Resonance Enhanced Covariance Matrix Detection

To avoid harmful interference to primary user system and improve the utilization of spectrum resource, it is necessary for the spectrum sensing algorithms of cognitive radio system to detect primary users’ signals and identify spectrum resource available as fast as possible under very low signal-to-noise (SNR) environments. Recently, covariance matrix based spectrum detection algorithm has been proposed as an efficient blind spectrum sensing method in cognitive radio context because it can avoid the noise uncertainty problem suffered by energy detection. To further improve the spectrum sensing performance under very low SNR region, a stochastic resonance enhanced covariance matrix based spectrum sensing algorithm is proposed in this paper. By adding specific optimal signal into the received signals, the proposed covariance matrix based spectrum sensing method can enlarge the deflection of the detection statistics’ probability density functions (PDF), which are associate with primary user signal existing, or not. Then, the spectrum sensing performance can be improved under very low SNR region. Comparing with existing covariance matrix based spectrum sensing algorithm, simulation results show that the proposed spectrum sensing method can significantly improve the detection probability as well as reduce detection period for the same probability of false alarm under very low SNR region.