Research On Wideband Spectrum Sensing From Compressed Sensing

Wireless spectrum is a necessary and non-renewable vital resource in today’s society. Takinginto account the growing demands of wireless communication services, it is necessary to have adeep study in how to effectively improve the spectrum utilization. Recently, the related research ofCognitive Radio (CR) has attracted much attention, it is possible to effectively use the spectrumholes to improve the spectrum utilization, and then alleviating spectrum resource scarcity problem.This capability of cognitive radio is realized by spectrum sensing. Spectrum sensing is defined as atechnique for achieving information about the spectral usage and existence of primary users (PUs).Future cognitive radio techniques will be capable of sensing a wideband of frequencies up toseveral GHz, for wideband spectrum sensing, the Nyquist sampling method is used to samplewideband signals, but the sampling rate is too high. This is a considerable challenge to implementwideband spectrum sensing.Recently, a new signal sampling method has emerged, called Compressed Sensing (CS)theory, it could be able to overcome the problem of high sampling rate. CS theory suggests that, ifthe signal is sufficiently sparse, it can be acquire signals at rates significantly lower than Nyquistsampling rate, the original signals can be recoverd form few sample values. In addition, due to thelow percentage of spectrum occupancy by active radios, wireless communication signals aretypically sparse in the frequency domain, allowing us to use compressed sensing to alleviate thehigher sampling rate problem.In the paper, the main research content and innovations as follows:Chapter one firstly introduces the background of our research in the paper, and have a detailedoverview of the basic theoretical knowledge, key technologies and research status at home andabroad of cognitive radio, then focuses on the analysis of cognitive radio spectrum sensingtechnologies.Chapter two studies relevant theoretical knowledge of compressed sensing technique, andintroduces detailly from three aspects: the sparse representation of the signal, the design ofmeasurement matrix and the dedign of reconstruction algorithms.Chapter three firstly detailed describes the important relevant reconstruction algorithm ofcompressed sensing. Then studies and discusses the main greedy pursuit algorithms. On this basis,contrapose the reconstruction of the unknown sparsity compressed sensing signals, a newcompressed sensing reconstruction algorithm is proposed in this chapter: Adaptive RegularizedSubspace Pursuit (ARSP) algorithm. Simulation results show that the algorithm is superior to other algorithms at its running time, complexity and reconstruction performance.Chapter four mainly researches in wideband spectrum sensing schemes. A widebandcooperative spectrum sensing scheme based on compressed sensing ARSP algorithm is proposed. Inthis scheme, multiple cognitive users take ARSP algorithm to reconstruct wideband signal under thesituation of unknown signal sparsity, and then to detect cooperately. Another wideband spectrumcooperative sensing scheme is proposed based on the reconstruction of the sub-channel energy.Firstly, the scheme reconstruct sub-channel energy vector. Secondly, the Generalized LikehoodRatio (GLR) test statistic is obtained by the reconstructed energy vector. And then to cooperatespectrum sensing. Simulation results show that the detection performance of these schemes aresuperior to conventional spectrum sensing methods, and multi-user cooperative spectrum sensingperformance is better than the single-user spectrum sensing.Chapter five summarizes the content of this paper, and points out the inadequacies of the paper,which require further study.