Research On Security Defense And Performance Optimization Of Cooperative Spectrum Sensing Under SSDF

With the rapid development of mobile communication technology and the rapid growth of wireless subscribers and mobile applications,substantial spectrum bands are occupied.The limited radio spectrum is overcrowded,and the scarcity of spectrum resources is increasingly serious.As a technology that can effectively improve the utilization of spectrum resources,cognitive radio(CR)came into being.In CR,secondary user(SU)can perceive the surrounding electromagnetic environment through spectrum sensing,to access spectrum resources(also known as spectrum holes)not occupied by primary user(PU).As key element of CR,spectrum sensing should be able to quickly and accurately detect the existence of spectrum holes,to avoid interference on PUs and enhance utilization of spectrum holes.By combining sensing information of multiple SUs,cooperative spectrum sensing(CSS)greatly improves the performance of spectrum sensing.However,the openness of CR also makes CSS face serious security problems.The attacker who launches spectrum sensing data falsification(SSDF),also known as malicious secondary user(MSU),leads CSS to make wrong decisions by maliciously falsifying its spectrum sensing data,thereby reducing the accuracy of CSS and deteriorating the performance of cognitive radio network(CRN).Aiming at the defense methods and performance optimization of CSS under SSDF,this thesis uses different levels of defense algorithms and performance optimization methods and builds a hierarchical and all-round defense network to improve the ability of CSS to defend SSDF and the performance of CRNs under SSDF.This research is to design the defense methods and performance optimization algorithm for the defending SSDF methods of CSS,the frame optimization of CSS under SSDF and sensing channel enhancement technology of CSS under SSDF.The effectiveness of the proposed algorithms under SSDF is verified by theoretical analysis and simulation.The research contents are detailed as follows.1.Aiming at the extreme performance degradation of existing reputation algorithms against massive SSDF,a reverse reputation algorithm defending against massive SSDF is proposed based on the deference of sensity to SSDF between independent spectrum sensing and CSS.While sensing results of SUs are reference,the reverse reputation value of the global fusion results is calculated to estimate the global sensing accuracy.The algorithm compares the global sensing accuracy and independent spectrum sensing results via a Generalized Likelihood Ratio Test(GLRT)based comparator,and SUs select the sensing results with higher accuracy as the final sensing decisions.Theoretical analysis and simulation results show that compared with the classical CSS algorithm,the proposed algorithm still has high sensing accuracy under massive SSDF.2.Aiming at the data fusion and defending SSDF method of CSS in the scene where SUs have inconsistent spectrum states,a CSS algorithm based on graph cut and spectrum sensing strategy selection is proposed.Firstly,the graph cut algorithm is applied to the data fusion.The spectrum state undirected graph is constructed based on the sensing information and topology information.The undirected graph is segmented by the graph cut method to obtain the spectrum decision of each SU.Furthermore,to defend against SSDF,a spectrum sensing strategy selection algorithm is proposed to compare the performance of data fusion results and independent sensing results,and select the sensing results with better performance as final decisions.Theoretical analysis and simulation results show that the proposed algorithm can effectively fuse the sensing information with inconsistent spectrum status,and sensing accuracy under SSDF is enhanced.3.In order to improve throughput of CRNs under SSDF,the framework optimization methds are studied.To avoid inevitable waste of time resources in the classical framework,a dual slots framework based on the joint optimization of sensing slot and sensing information transmission time slot is proposed.After establishing the non-convex optimization problem to optimize throughput,a dual slots framework optimization algorithm based on succesive convex approximation(SCA)is proposed to obtain the solution of the non-convex problem,and the performance of the proposed algorithm is verified by simulation.Based on the researches of framework optimization without SSDF and defending SSDF methods,for framework optimization under SSDF,a double reputation values framework and an optimization algorithm based on alternative optimization(AO)is proposed.By calculating the double reputation values of SUs respectively,the equivalent probability of detection and equivalent probability of false alarm are derived.Then,the throughput optimization problem is established.The optimization parameters include double reputation thresholds,sensing time,sensing threshold and so on.For the non-convex optimization problem,a double reputation values framework parameters optimization algorithm based on AO is proposed to obtain the solution of the original problem.Theoretical analysis and experimental results show that the proposed framework and parameters optimization algorithm can effectively improve throughput of CRNs under SSDF compared with the fixed parameters framework.4.Aiming at improving the performance of CSS under SSDF,the sensing channel enhancement methods are studied with assistance of intelligent correlation surface(IRS).For IRS assisted CSS,the optimization problem of joint beamforming and framework parameters optimization is established.For this complex non-convex problem,the equivalent optimization problem is obtained by using Ky Fan norm and other methods.To solve the equivalent problem,a joint beamforming and framework parameters optimization algorithm based on SCA is proposed,and the effectiveness of the algorithm is verified by simulation.Based on the researches of IRS assited CSS without SSDF and defending SSDF algorithms,an IRS assisted reverse reputation value algorithm(IRS-RR)is proposed to defend SSDF.The sensing performance of trusted users and the performance of global results are compared by reverse reputation value,and the sensing accuracy optimization problem is established.For the complex non-convex optimization problem,the equivalent optimization problem is obtained,and the IRS-RR parameter optimization algorithm based on AO is proposed to solve the equivalent problem and obtain the solution of the original non-convex problem.Theoretical analysis and simulation results show that compared with the fixed parameter algorithm,the proposed algorithm can improve the the sensing channel and enhance the sensing accuracy.