Research On Defense Algorithm Of Internal And External Collusive Attack In Cooperative Spectrum Sensing

The advent of the 5G epoch makes the shortage of spectrum resources more severe.Cooperative Spectrum Sensing(CSS)in cognitive radio is considered to be the most effective way to improve the utilization of spectrum resources.However,malicious users take the advantages of the openness of their network channels can initiate Spectrum Sensing Data Falsification Attack(SSDF)attack at any time,and disturb the judgment result of the fused sensing data.The introduction of trust evaluation mechanism can resist this perceived threat effiectively,but the vulnerability of the trust evaluation mechanism also provides malicious users with new attack chance.The paper deeply analyzes the two role attributes that the sensing user can engage in CSS activities: Initiating Sensing User(ISU)and Cooperating Sensing User CSU.The study found that the current trust evaluation mechanism existence the loopholes for reviewing ISU feedback data is lack after the end of CSS activities.Malicious users can be organized and premeditated,then implement Collusive False Feedback(CFF)attacks through internal and external collusion to obtain high sensing trust value,thus bypasses the detection of the trust evaluation mechanism.In this regard,the paper proposes a two-layer defense mechanism of CFF attack-FeedGuard,from the perspective of feedback trust and I-C frequency correlation analysis.The main research work and innovations are as follows:· Describle the CFF attack strategy and attack process detaily.CFF attackerscan raise the sensing trust value between each other through internal and external collusion.Specifically,a CFF attacker initiates a CSS activity as an ISU,and after the end of the activity sends feedback data that is beneficial to his conspirators to the data fusion center,thereby implements SSDF attack more effectively under the cover of high-sensing trust values to achieve the purpose of disrupting the final decision of the fusion center.The simulation results show that compared with the traditional malicious users' self-effort mode and intermittent dynamic SSDF attack behavior,CFF attacks can quickly raise the sensing trust value of malicious users in batches and generate more error of sensing trust value in the network value.· Introduced the concept of feedback trust value and related calculation methods to quickly detect whether the ISU requesting CSS activity is a CFF attacker,and construct the first layer of FeedGuard defense mechanism,and stop the CFF attack behavior that will be initiated with the low feedback trust value from the source.At the same time,analyze the historical feedback data of the ISU and the historical sensing data of the CSU to calculate the I-C frequency exhibited by the internal and external collusion features,and the second layer of the FeedGuard defense mechanism is constructed to detect the CFF attacker hidden in the CSU.According to the detection result of the second layer defense mechanism,perfect the calculation method of the sensing trust value,so that it is difficult for the CFF attacker to obtain the high perceived trust value.The simulation results show that the FeedGuard defense mechanism can reduce the sensing trust value error generated by the CFF attacker in the network effectively and suppress the generation of malicious response times.· Integrate the FeedGuard defense mechanism into the cooperative spectrum sensing data fusion algorithm to form a hard-threshold fusion resist CFF attacks to protect the hard-threshold fusion criteria of “OR”,“AND” and “Majority”,and resist CFF attacks soft-threshold fusion criterion,The decision fusion criterion is used to improve the classical SPRT algorithm,and obtain the Trusted Sequential Probability Ratio Detection Algorithm(FSPRT).the flow of detailed hard decision protection and soft decision protection algorithm implementation diagram is given.The simulation results show that the FeedGuard defense mechanism can not only suppress the attack success rate of CFF attacks for the "OR","AND" and "Majority" hard decision fusion criteria,but also effectively improve the sensing performance advantage of the SPRT soft decision fusion algorithm in the face of CFF attacks.