Research On Key Technologies Of Cooperative Spectrum Sensing In Cognitive Radio Networks

The emergence of cognitive radio technology has brought hope to solve the problem ofspectrum shortage. It enables cognitive radio users to access the idle spectrum in an “opportunis-tic way” without causing exploit interference to the primary users, resulting in an improvementof the spectrum utilization. As the prerequisite and foundation for realization of cognitive radio,spectrum sensing technology has been drawing more and more attention of academic area. Co-operative spectrum sensing overcomes the limitations of single-user local spectrum sensing, butadditional overhead is introduced inevitable while improving the sensing performance, whichneeds to be considered. This dissertation focuses on the overhead of conventional cooperativespectrum sensing scheme based on the energy detection algorithm. The main contributions areoutlined as follows:1. A double-threshold cooperative spectrum sensing algorithm based on noise uncertainty isproposed, which takes the contradiction between the joint detection performance improvementand the control channel burden into account. In this algorithm, two decision thresholds are cal-culated according to the noise power uncertainty, which enhances the robustness of the system.Considering the fact that the local sensing result of cognitive radio users between the two thre-sholds still has offers valuable information, a adaptive sequential step combination algorithmbased on equal gain combining method is also been put forward, which aimed at further advanc-ing the joint sensing performance. Simulation results show that the proposed algorithm obtainbetter performance with a lower communication overhead, while increase slightly the complexityof the control system.2. Aiming at the problem that the detection time maybe too long and the overhead of thecontrol channel of traditional sequential detection in the case of low signal-to-noise ratio may betoo large, a cooperative spectrum sensing algorithm based on truncated sequential probabilityradio test is presented. A split-phase processing method is employed firstly; all the CR nodes in-volved quantities its local energy-based sensing samples and compared with specified thresholdconsequently. Instead of sending all the sensing results of cooperative CR user, only the numberof sampling points bigger than the threshold is sent to fusion center serving as local detectionstatistics, while sequential test is employed at the fusion center. Gaussian distribution is approx-imately obtained based on central limit theorem and DeMoivre-Laplace theorem while the sig-nal-to-noise ratio of cognitive radio users are same and not, deriving process of likelihood ratiofunction is reasonably simplified. Theoretical analysis and simulation results indicate that, com-pared with existing energy detection method, the proposed algorithm reduces the number of sapling points largely in the condition of low signal-to-noise ratio.3. A cooperative spectrum sensing algorithm is given based on the on-off model of licensedband occupancy and Bayesian decision rule. In practice, the moment that each user initiatesspectrum sensing, sampling and transmitting local sensing information to the fusion center maybe asynchronous, what’s more, the behavior of primary user affect the sensing performance ofcognitive radio system. The proposed algorithm takes all these factors into account, and it con-clusion could be applied to both synchronous and asynchronous situations. Theoretical analysisand simulation results show that, the better performance is achieved with the proposed algorithm,while compared with the synchronous cooperative spectrum sensing scheme.4. A test platform based on MATLAB GUI is designed, which can be used to evaluate prac-tical performance of proposed cooperative spectrum sensing algorithm. The architecture of thetest platform and the elaborate design of each module are presented. The performance assess-ment of equal gain combining algorithm is functioned on the platform. Testing results keeps con-sistent with theoretical analysis, which verifies the reliability and practicality of the platform.