The Cognitive Radio Spectrum Perception Algorithm Based On Compression Perception Research

Compressed sensing has a great potential in the cognitive radio (CR) system. This thesis is supported by National Science Funds and attempts to make a contribution to the theory and application of CR system.In this paper, the basic principle of compressed sensing technology and its application of CR spectrum sensing procedure are investigated. Several creative algorithms are developed in this thesis.This paper addresses the problem of designing an appropriate quiet period management scheme over time-variant channels. To achieve better performance over time-variant channels, we propose a flexible quiet period management scheme which can adjust the sensing parameters according to the previous sensing result and the information of ACK. The performance of the proposed scheme has been evaluated through Markov analysis. Numerical results show that under time-variant channel conditions, a better probability of detection and higher channel utilization is achieved compared to the traditional fixed quiet period management scheme.Collaborative spectrum sensing (CSS) can significantly improve the performance of spectrum sensing based on the spatial diversity gain of different cognitive radio. In wideband spectrum sensing scenario, since there might not be enough CRs in the network, or due to hardware limitations, each CR node can only sense a relatively narrow band of radio spectrum. Consequently, the available channel sensing information is far from being sufficient for precisely recognizing the wide range of unoccupied channels. Based on the fact that the spectrum usage information the CR nodes collect has a common sparsity pattern, in this paper, we present a compressed collaborative wideband spectrum sensing scheme in cognitive radio networks. Under the hypothesis of joint sparsity, the CRs need to randomly detect a very small number of sub-channels according to a measurement matrix and send the results to a fusion center. To make the compressed sensing more effective, the scheme uses LDPC-like measurement matrix. Then the whole channel status can be recoverd by the fusion center through a low-complexity message passing algorithm. Numerical results show that under a joint sparsity model, using the proposed distributed compressed sensing scheme, the CRs make a small number of measurements and get a high probability of detection.In order to reduce the sampling rate in the broadband spectrum sensing, a cognitive radio spectrum sensing algorithm based on distributed compressive sensing is proposed. In IEEE802.22standard, Timeline is divided into successive superframes. Each superframe consists of a number of frames, and quiet period is set in each frame. At the end of each superframe, the spectrum sensing results of the quiet periods in this superframe are merged. Because the primaiy user signals are sparse in the frequency domain, compressive sensing method can be adopt. Assume that the primary user varys slowly, then the sampling sighals in the same superframe are joint sparse and meet the joint sparsity model JSM-2. So the second user makes random measurements below the Nyquist sampling rate using the measurement matrix. After that, using DCS-SOMP algorithm, the data of each sample of the quiet periods can be reconstructed. Merge the reconstructed data, then the reasonable judgment can be made.A CS-Feature detection spectrum sensing algorithms for cognitive radio is proposed in this paper. The traditional feature detection algorithm based on cyclic spectrum density has a very high accuracy, but it can’t be widely used because of the high complexity and very long detection period. The CS-Feature detection in this paper is designed based on the sparsity of the cyclic autocorrelation. Because most of the man-made signals are cyclostationary, the values in cyclic autocorrelation domain are sparse. From the measurements based on the compressed sensing of the cyclic autocorrelation, the actual cyclic autocorrelation of the signal can be recovered. From the simulation, the dissertation can be made that a very simple OMP algorithm can get a high probability of detection.Finally, the content of the whole dissertation is summarized, and several valuable research directions of compressed sensing and CR spectrum sensing are discussed.