Research On Spectrum Sensing Algorithm In Cognitive Radio Network

Cognitive radio technology provides the best solution to solve the spectrum tense and the raising of utilization ratio of spectrum, while the spectrum sensing technology is the foundation of cognitive radio technology, which performance is directly related with the performance of the system, so it is of great significance to do research on this subject. Spectrum sensing in a real environment is vulnerable to noise power, low SNR condition such as volatility etc. And in broadband spectrum perception, due to the increased communication bandwidth, the too high rate of ADC sampling in the process of problem, spectrum sensing needs more and more high requirement of sensing device. In order to solve the spectrum sensing problem of uncertain noise variance, low signal-to-noise ratio, and broadband communication under the condition of different spectrum sensing, in this paper, the perception algorithm and cognitive strategies are studied, the main work is as follows:Firstly, in view of the noise power fluctuations within a certain range, the noise variance under the condition of uncertainty energy sensing algorithm, based on the analysis of the cognitive performance, this paper puts forward the noise double threshold energy empirical sensing algorithm. For different SNR, adaptive changes the spectrum sensing time, comparing with the traditional energy sensing algorithm,the proposed algorithm has high detection performance and good anti-jamming ability. On this basis, in order to better improve the detection of the system, in this paper, the secondary energy efficiency optimization problems of cooperative spectrum sensing based on double threshold are studied, putting forward the optimization algorithm to improve the efficiency of cooperative spectrum sensing,through the simulation verifying the effectiveness and the superiority of the proposed optimization algorithm.Secondly, in view of the low signal-to-noise ratio spectrum perception, the research of spectrum sensing algorithm was carried out based on stochastic resonance technology. According to the actual low SNR environment, based on that stochastic resonance system has good properties to enhance the system output of signal-to-noise ratio, the principle of stochastic resonance is introduced into thetraditional spectrum perception energy detection method, this paper proposes one energy spectrum sensing algorithm under the condition of low signal-to-noise ratio based on stochastic resonance technology, the detection probability theoretical expression of the method was deduced, the simulation shows that this algorithm effectively improves the detection performance of the system. On this basis,according to the advantage of double threshold algorithm, the two-step spectrum sensing algorithm based on stochastic resonance was proposed,different to the traditional stochastic resonance spectrum sensing algorithm, this algorithm, changes the stochastic resonance system parameters with low SNR environment by setting the best two-door limit, so as to improve the algorithm adaptability and flexibility.This article dose optimization to the performance objective function on the two aspects of the maximization of cognitive network throughput and the minimality of perception time, This paper proposes a fast 2 d optimization iteration algorithm,through the simulation, the proposed algorithm effectively improve the efficiency of the spectrum perception and the detection performance of the system.Finally, in view of the broadband spectrum sensing, this paper proposes a high efficient spectrum sensing method based on compression sensing. In cognitive radio networks, due to the different cognitive users and the authorized user in different space, the difference of the distribution of the environment, and the complexity of the channel, they may have different sparse spectrum, which is not suitable for information sharing in the process of joint spectrum sensing. By using each CR nodes with compressing sampling technique for spectrum estimation to reduce the sampling rate and cost, through the Bayesian model to seek the optimal parameters,completing spectrum information detection, and use the single layer expand collaborative spectrum sensing algorithm to put forward using the non-parametric information grouping mechanism, and realize automatic grouping by introducing Dirichlet process multi-layer to Bayesian model,as well as the compressed sensing data and reasoning of Shared parameters, and passed by super parameter to choose the best fusion center for spectrum, eventually being passed to SU. The algorithm makes full use of different layers of compressed sensing data each time when a user acquired and does collaboration for spectrum sensing to fusion, thus effectively improves the performance of spectrum sensing.