Research On Multi-node Cooperative Spoofing And Anti-spoofing In Spectrum Sensing

Cognitive radio is widely regarded as an effective method to improve the utilization rate of spectrum and solve the shortage of spectrum resources by allowing dynamic spectrum access.Spectrum sensing is the cornerstone of cognitive radio,which provides guidance for dynamic spectrum access.Due to the openness of the wireless communication environment and the particularity of the cognitive radio,the spectrum sensing process is facing potential security problems.In the sensing environment where noise and channel fading exist,multi-node cooperative spectrum sensing can effectively improve spectrum sensing performance compared to single-node spectrum sensing.However,the multi-node cooperative spectrum sensing process is susceptible to malicious users,which confuse the cognitive system to make an unreliable sensing result,resulting in a degration of collaborative spectrum sensing performance.Based on the analysis of existing spectrum sensing spoofing and anti-spoofing,this thesis mainly research the multi-node spoofing and anti-spoofing in spectrum sensing process.This thesis analyzes the influence of the multi-node cooperative spoofing on the spectrum sensing performance,and then proposes a multi-node cooperative spoofing manchanism.The proposed multi-node cooperative spoofing manchanism is used to learn the relevant parameters of the sensing network through multi-node cooperation,and the corresponding spoofing strategy is selected to attack the cognitive network.This thesis analyzes the learning length and the learning accuracy of the malicious user in learning stage,and studies the specific multi-node cooperative spoofing method.Finally,the validity of the multi-node cooperative spoofing manchanism is verified by computer simulations.A method of multi-node cooperative spectrum sensing based on clustering is proposed to reduce the adverse effects of malicious users on the cognitive system.The algorithm uses the clustering method to identify the contaminated sensing data without any priori information on both the attackers and the cognitive users,and then assigns an identity value to each cognitive user to reduce the clustering identification errors.Finally,both of the robustness and validity of the proposed algorithm are verified by the simulation in the case that the sensing link and the cooperative sensing stage are disturbed by malicious users.