Adaptive Bistable Stochastic Resonance Aided Spectrum Sensing

As a fundamental technology of cognitive radio, the spectrum sensing scheme is required to perform well in low signal-to-noise ratio (SNR) environments. In this paper, we propose a novel spectrum sensing method based on adaptive bistable stochastic resonance (A-BSR). By maximizing the SNR gain introduced by the BSR system, we first present an A-BSR system, of which the parameters can be adaptively adjusted based on the background noise. Then, we propose an A-BSR aided spectrum sensing scheme by passing the received signal through an A-BSR system to improve SNR. Based on the characteristics of A-BSR system output in frequency and time domains, we further propose two energy detection (ED)-based spectrum sensing algorithms. As the output of an A-BSR system given a noise input is concentrated around frequency zero, we propose a modified ED based on periodogram (P-ED) in frequency domain. Moreover, as the A-BSR system output for noise input has approximate constant amplitude in time domain, a novel deviation-based ED (D-ED) is proposed. Extensive simulation results show that the proposed A-BSR aided spectrum sensing scheme can achieve much better performance than the existing ED and BSR aided spectrum sensing schemes with fixed parameters, especially under very low SNR region.