Research And Application Of Spectrum Sensing Approach Based On Stochastic Resonance In Cognitive Radio Networks

Spectrum sensing, one of the key technologies of cognitive radio, needs to make a fast and accurate detection of idle spectrum around to avoid the interference to primary users. But the low signal-to-noise ratio (SNR) environment will cause a great impact on the accuracy of detection results. Stochastic resonance phenomenon described in a nonlinear system, when there is a match between the input signal and noise, the noise energy will be transferred to the signal energy, the SNR of the system output will not be lowered, but will increase. it is very suitable for the detection of weak signals. In order to enhance the accuracy of idle band detection under the low SNR environment, stochastic resonance system is introduced to the traditional energy detection in this paper, and we studied the energy detection based on stochastic resonance in cognitive radio.Firstly, the research background and significance is described in this paper, the research status of spectrum sensing in cognitive radio and the research and application of stochastic resonance in signal detection and the research status of spectrum sensing based on stochastic resonance in cognitive radio are also introduced respectively. Secondly, we make a classification to the existing spectrum sensing technologies, it is divided into single-node spectrum sensing and multi-node cooperative spectrum sensing according to the number of cognitive users. This thesis focuses on principle and algorithms of energy detection and cyclostationary feature detection and matched filter detection in single-node spectrum sensing. In addition, three kinds of hard information-combining are be introduced in this paper, there are "OR" rule,"AND" rule and "K rank" rule.Then the principles of traditional bistable stochastic resonance is introduced, the detection performance of energy detection based on traditional bistable stochastic resonance and traditional energy detection both are studied, formulae describing the detection probability and the false alarm probability of two energy detection are also derived. After the basic steps of the simulation experiments are described, this paper compared the performance of the two energy detection by Matlab. The experiments show that the detection performance of energy detection based on traditional bistable stochastic resonance is better than traditional energy detection.Next, a novel monostable stochastic resonance is introduced into the energy detection. First, the basic principles of this monostable stochastic resonance is introduced, then we compared the performance of the novel energy detection and traditional energy detection by Matlab. The experiments show that the detection performance of energy detection based on novel monostable stochastic resonance is better than traditional energy detection, but it’s detection performance is influenced by signal frequency.Finally, this thesis describes the basic principles of a series of single-mode nonlinear optical system, focus on the application of these types of stochastic resonance in the traditional energy detection, compared the performance of energy detection based on these types of stochastic resonance and traditional energy detection by Matlab. The experiments show that all the detection performance of energy detection based on these types of stochastic resonance are better than traditional energy detection, but their detection performance are influenced by signal frequency also.