Spectrum sensing in low SNR regime via stochastic resonance

Kun Zheng,Husheng Li,Seddik M. Djouadi,Jun Wang

DOI: https://doi.org/10.1109/CISS.2010.5464907

2010-01-01

Abstract:Spectrum sensing is essential in cognitive radio to enable dynamic spectrum access. In many scenarios, primary user signal must be detected reliably in low signal-to-noise ratio (SNR) regime under required sensing time. We propose to use stochastic resonance, a nonlinear filter having certain resonance frequency, to detect primary users when the SNR is very low. Both block and sequential detection schemes are studied. Simulation results show that, under the required false alarm rate, both detection probability and average detection delay can be substantially improved. A few implementation issues are also discussed.