Research On Spectrum Sensing Technology In Different Wireless Environments

With the rapid development of radio technology and communications services,the demand for wireless communication at anytime and anywhere is increasing.At the same time,the demand for radio spectrum resources from different services and sectors shows an explosive growth.The contradiction between the supply and demand of limited spectrum resources is increasingly tense,and the contradiction of structural shortage is prominent.However,the survey data shows that average occupied ratios of authorized spectrum in multiple dimensions(time,frequency,spatial and energy domain)are low and there is a lot of idle spectrum with low utilization rate.Cognitive radio technology can 'opportunity'access to idle spectrum,effectively improve the utilization rate of spectrum resources,and has drawn extensive attention.As a key technique for the cognitive radio,the spectrum sensing is also a fundamental element in the cognitive radio system to share spectrum.Only a reliable spectrum sensing technique could find out idle spectrum and select appropriate frequency bands for secondary users in real-time.However,the spectrum sensing performance is closely related to the complex and changeable external electromagnetic environment.In order to discover idle frequencies in multiple dimensions space in real-time,it is necessary to have a high spectrum sensing efficiency.And also in order to avoid causing harmful interference to primary users,the spectrum sensing is required to be highly reliable.This paper makes an in-depth research on the influence of different wireless environment on spectrum sensing performance.The main work and contributions can be summarized as follows:1)Based on the spectrum sensing system model,this paper analyses the conventional spectrum sensing algorithms,namely Energy Detector,Matched Filter Detector and Generalized Likelihood-Ratio Detector algorithm,as well as the cooperative spectrum sensing algorithm and then concludes their merits and demerits.In Gaussian noise ideal channel environment,the sensing performance of each algorithm can reach its best.In non-Gaussian noise fading channel environment,the sensing performance of each algorithm decline with varying degrees.2)For the 'hidden terminal' problem,a multiple cluster cooperative wideband spectrum sensing algorithm is proposed and the performance analysis is conducted.First,the proposed method utilizes multiple clusters to sense different narrow band spectrum separately over Gaussian noise small scale fading channels.Each cluster head uses OR rule to fuse and judge the information of the secondary users.Thus the cooperative spectrum sensing in the cluster is completed.Secondly,each cluster head transmits decisions to the fusion center over small or large scale fading channels.Finally,the fusion center converges on the decision results of each narrow band spectrum,and realizes the spectrum sensing of the wideband spectrum in the limited area.At the meantime,an analysis of the number of secondary users involved in spectrum sensing in clusters and the results show that when a certain number of secondary users is reached,the proposed method does not increase the communication load for the cognitive network,Furthermore,it could reduce the sensing process and energy consumption of the rest of secondary users.3)In view of the effect of non-Gaussian noise on spectrum sensing performance,a spectrum sensing algorithm based on the maximum generalized correntropy is proposed.Using the maximum correntropy criterion for the robustness of the non-Gaussian noise abnormality,a system model based on the generalized Gaussian kernel function is proposed in SaS noise.First,the proposed method utilizes multiple antennas to form the received signal matrix.Secondly,it assigns different weight vectors to different received signals with the help of signal sparse representation.Thirdly,by using the convex optimization technique,the non-negative optimal solution of the system model is obtained,that is,the optimal weight vector.Thus,the spectrum sensing in SaS noise environment is realized,and the spectrum sensing performance under low generalized signal to noise ratio(GSNR)is improved.Finally,the simulation verifies the correctness,convergence and complexity of the theoretical analysis of the proposed algorithm.The results show that the algorithm proposed in this paper can effectively solve the spectrum sensing problem in SaS noise environment,and improve the spectrum sensing performance in the low GSNR environment.4)For the issue of spectrum sensing in non-Gaussian noise multipath fading channel environment,a spectrum sensing algorithm based on Rao detection is proposed.A spectrum sensing model which is used in non-Gaussian noise multipath fading channel environment is designed in this paper.First,when the PDF of non-Gaussian noise has a closed-form expression and multipath fading channel has unknown amplitude fading parameters,spectrum sensing algorithm based on Rao detection is directly adopted to realize the spectrum sensing.Secondly,when the PDF of non-Gaussian noise does not have a closed-form expression and multipath fading channel has unknown amplitude fading parameters,the PDF of non-Gaussian noise is estimated approximately by using the nonparametric window estimation method,and spectrum sensing algorithm based on Rao detection is used to realize the spectrum sensing.Finally,the simulation is carried out in different non-Gaussian noise and different multipath fading channels.The theoretical analysis and simulation results show that the proposed algorithm can realize better spectrum sensing in the non-Gaussian multipath fading channel environment.