Research Of Cooperative Spectrum Sensing Technology Based On Fusion Algorithm

Under the traditional fixed spectrum allocation policy, the utilization of global spectrumresources shows a sign of high degree imbalance. In order to deal with the contradiction betweeninsufficient utilization of licensed spectrum and the shortage of spectrum resources, CognitiveRadio (CR) was proposed. As an intelligent spectrum sharing technology, CR can effectivelyimprove the spectrum utilization. Spectrum sensing (SS) is one of the key technologies of CR, andalso an important premise for practical application of CR. In this dissertation, the spectrum sensingalgorithms are thoroughly investigated.The concept, key techniques of cognitive radio, and its domestic and international researchsituation, and its application prospects are described in details. This dissertation analyzes thedetailed classification of spectrum sensing algorithms. In single-user spectrum sensing technologies,we focus on the analysis of energy detection, matched filtering detection and cyclostationary featuredetection based on transmitter detection; in multi-users cooperative spectrum sensing technologies,we focus on the analysis of data fusion rules based on hard decision and soft decision.In order to solve the problems that conventional multi-users cooperative spectrum sensingalgorithm ignores the environmental differences of cognitive users and that the application of D-Sevidence theory neglects conflicts between evidences, a spectrum sensing algorithm based onimproved D-S evidence theory is proposed in this dissertation. Under the condition withoutchanging combination rule, improving D-S evidence theory through similary functions. A detaileddescription of the algorithm’s process about how to apply the improved D-S evidence theory inspectrum sensing is given. Simulation analysis proves the performance advantages of the algorithm.Because of the effects of the shadowing, multipath fading and noise uncertainty, when energydetection is used in local spectrum sensing, the sensing result may be incorrect. To deal with thissituation, a cooperative spectrum algorithm with double-thresholds based on noise uncertainty isproposed in this dissertation. Double-thresholds energy detection is selected in local sensing; thefusion center employs different methods according to different forms of sensing information. Thetheoretical analysis and simulation results both show that the algorithm can improve the detectionperformance significantly.