Research On Cooperative Spectrum Sensing Technique In Cognitive Radio

Cognitive radio was proposed by Doctor Joseph Mitola III in his PhD. dissertation in 1999. The spectrum could be allocated without changing the existing spectrum allocation, which can improve the ratio of the spectrum. Spectrum sensing is the fundation of the cognitive radio. Cognitive radio offers a variety of research fields, including spectrum sensing, spectrum sharing, spectrum management and power allocation. Now, in cognitive radio, the spectrum sensing based on the detection of the primary user can be divided into three categories: matched filter detector, energy detector, and cyclostationary feature detector. In this thesis, the spectrum sensing was studied deeply.In this thesis, the spectrum sensing based on traditional energy detection and entropy detection algorithm was studied. The traditional energy detection algorithm has the problem of low probobality of detection and sensitive to noise uncertainty. So a novel energy detection algorithm, which could improve the probobality of detection and counteract the sensitive to noise uncertainty, was proposed. The simulation results, which are got from the Matlab, show that the novel energy detection algorithm improves the probability of detection and reduces the sensitive of noise uncertainty.Then, we studied the spectrum sensing algorithm based on entropy detection, which was proposed in 2009. Entropy is a parameter of description the degree of disorder. The entropy of the noise is larger than the signal'. The previous studies had proved that the entropy of noise is a constant, which intrinsically robust against noise uncertainty. In this thesis, we proposed a novel entropy detection algorithm in frequency domain, which reduced the calculation. Moreover, the detection probability and the sensitivity of noise uncertainty of traditional energy detection, enhanced energy detection and entropy detection algorithm were studied. The simulation results showed that the detection ability based on novel energy detection algorithm was the best, while the sensitivity of noise uncertainty of novel energy detection was the same as the spectrum sensing based on the entropy detection algorithm. The two algorithms were better than the traditional energy detection. In the last part, we give the simulation results of the cooperative spectrum sensing based on the traditional energy detection and entropy detection algorithm. The simulation results show that the cooperative spectrum sensing algorithm outperforms single entropy detection at least 5dB, when the probobality of detection is larger than 0.9 and the detection number is eight.