Research On Approaches Of Spectrum Sensing In The Presence Of Byzantine Attack For Cognitive Radio

With the rapid development of mobile communication technology,the demand for spectrum resources is increasing exponentially.The conflict between the soaring require-ment of spectrum and the limited spectrum resources has reached a plateau of radio devel-opment.However,large amount of spectrum resources waste in the traditional spectrum allocation method has lead to the general contradiction between the sharp increase in the demand for spectrum resources and the low utilization rate of spectrum resources.Cog-nitive radio is the core technology to solve this general contradiction effectively,and its primary task is spectrum sensing.However,it is urgent to solve the security problem in the process of spectrum sensing,especially for all kinds of attack behaviors represented by Byzantine attack.For the challenge of the Byzantine attack on cognitive radio spectrum sensing,this paper explores a algorithm model from the perspective of the integration of radio communication,machine learning,decision theory and uncertainty theory.The main work of this thesis is summarized as follows:(1)To solve the problem of spectrum sensing in the all-zero Byzantine attack envi-ronment,this paper propose a cooperative spectrum sensing model based on improved k-means clustering algorithm and evidential reasoning theory.Firstly,the improved k-means clustering algorithm is used to sensing the classification of attacking users,and the user weight is further quantified based on the clustering results and signal to noise ratio,so as to effectively avoid the impact of attacking users on the sensing results.Secondly,we use evidential reasoning theory to realize data fusion.Finally,the spectrum decision result is obtained by approximating ideal solution sort method and a preset threshold.Simulation results show that the proposed model has better perceptual performance than traditional perceptual algorithms.(2)A spectrum sensing model based on hierarchical clustering and evidential rea-soning is put forward to tackle the spectrum sensing problem of all-one Byzantine attack and anti-byzantine attack.Firstly,the attacking users are identified effectively by hi-erarchical clustering algorithm,and the user weight is quantified by clustering results.Secondly,the sensing evidence of each user is obtained by an energy detection algorithm,and the result of cooperative spectrum sensing is obtained based on evidential reasoning theory.Finally,the spectrum decision is made by comparing the fusion results with the pre-determined threshold.The simulation results show that compared with the traditional algorithm,this model can effectively improve the cognitive radio sensing performance in the all-one Byzantine attack or anti-byzantine attack environment.