Based On The Research Of Collaborative Spectrum Detection Algorithm For Cognitive Radio

As radio resource is becoming increasingly scarce, more andmore people realize the contradiction between the allocation and theusage of spectrum. The appearrance of Cognitive Radio will provideeffective solution to the problem of spectrum shortage. Spectrumsensing is one of most critical technologies in cognitive radio. Therefore,the sensing about physical layer plays an important role in CR.The thesis analyzes the concept of Cognitive Radio technology,the study status quo at home and abroad, and the spectrum sensingtechnology which are usually used, and focus on the principles andperformance of energy detection, laying a theoretical foundation forresearch. During the spectrum sensing, performance of single-nodedetection will degrade due to fading channel, the shadowing effectenvironment, so we introduce cooperating detection to the detectionprocess.In view of traditional detection, all the secondary users shouldparticipate to detect in collaboration. So on one hand it consume morebandwidth limited, the other hand it may appears serious false alarm orleakage alarm with deep fading of second users. These error messagesin collaboration also will affect the overall detection performance, so amethod of cooperative spectrum sensing based on threshold and SNRselected is proposed, use of threshold to do algorithm selection for SNR.At the same time, in order to further the effective secondary userselection, give full play to the advantages of double threshold, usingdetection energy of decision domain to improve the overall systemperformance, so an improved decision domain check double thresholdmulti-users collaborative spectrum detection algorithm is researchedand proved. And on the base of the detection performance ofsimulation experiments under more typical circumstances, when increase the probability of non-judgment region, the method willappears to detect false positive.In order to reduce the noise in the fading channel uncertaintyon the impact of energy detection method, in the tradition signaldetection, based on the covariance of signal detection and extractionmethod, this paper proposed an improved method based on theminimization of detection error of the autocorrelation matrix ofcooperative spectrum detection. This method does not need to knowthe user prior information as energy detection, and has low complexity,compared with the traditional energy detection, no need noisevariance while improves the detection performance, the detectionerror is minimized by the threshold optimization, it reducing interferenceprobability with primary user and increasing the utilization of idlespectrum. Computer simulation results show that when SNR=-18dB,based on this chapter program collaborative detection probability hasreached0.98, while ED(energy detection) scheme is0.76, with SNRincreased,based on autocorrelation matrix cooperative effect isoverall higher than ED scheme.