Research On Anti-noise And Anti-attack Performance Of Cognitive Radio Networks Based On Wide-band Cooperative Spectrum Sensing

In cognitive radio networks, spectrum sensing is one of the core technologies. Spectrum sensing can detect the presence of primary users to ensure that cognitive users are able to exploit spectrum holes. Due to the limitation of Nyquist sampling theorem, it is difficult for existing hardware to realize spectrum sensing in wideband environment with Nyquist sampling. Therefore, researchers introduce compression sensing into wideband spectrum sensing to reduce the sampling rate. At the same time, because of hardware limitation and the wireless environment, a single user can't accurately detect the spectrum. Multiple users can improve the detection performance and shorten the detection time due to the spatial gain. Cooperative spectrum sensing also brings some problems,such as the exchange of information will consume additional energy and time, and cooperative work will suffer from malicious users.In this paper, based on wideband compressive sensing, the anti-noise and anti-attack performance in wideband spectrum sensing is studied deeply. By introducing some prior information and machine learning algorithms, a robust wideband spectrum sensing algorithm is proposed.The main innovations of this paper are as follows:1.We propose a weight-based Bayesian compressive sensing fusion rule to resist noise. Bayesian compressive sensing algorithm has obvious advantages in resisting noise. Based on the traditional soft-decision model and cooperative Bayesian compressive sensing, this paper proposes a weighted based fusion rule. In addition to the traditional equal gain combination and maximal ratio combination, this paper propose the combination based on historical data and "error bars", which is based on Bayesian compressive sensing.2.We propose an angle-based malicious user detection to resist attacks. In spectrum sensing, there is a kind of malicious user that sends tampering result to the fusion center to interfere with the normal spectrum sensing, threatening the security of cognitive radio. This attack is called spectrum sensing data falsification attack. Based on the study of attack and defense behavior in wideband environment, this paper proposes an angle-based outlier detection method. Compared with the traditional distance-based malicious user detection, the angle-based method has a better performance in the wide-band environment. The simulation results show that the proposed method can detect the existence of independent malicious users.3.We propose a density-based malicious user detection to resist attacks. In spectrum sensing data falsification attack, malicious users can cooperate to pose a greater threat to the cognitive radio networks. In this paper, we propose a density-based attack detection method to resist independent attack and dependent attack. This method is based on density clustering, and is compatible with distributed compressive sensing. The simulation results show that the method proposed in this paper can detect the existence of independent malicious users and cooperative malicious users.