Research On Compressive Spectrum Sensing And Spectrum Sharing Transmission Technology In Wirless Communication

With the development of wireless communication, spectrum resource became moreand more tightly. But, currently methods of spectrum division can not handle thisdilemma. Thus, Cognitive Radio as a smart spectrum sharing system had been proposed.It can reuse the spectrum resource and prompote the uitilization of spectrum, which hasbeen drawn great attentation in recently years. Spectrum sensing technology is aprecondition for the realization of Cognitive Radio, but currently sensing methods cannot be applied in broad band detection due to the restriction of Nyquist samplingfrequency. As a recently popular information processing technology, compress sensinghas ability to break the constraint of Nyquist sampling frequency. Thus, it has beenaroused great interests once it has been proposed. Now, more and more researchers havepaid attentation to this technology, including theoretical research and application inother relevant fields like Spectrum Sensing. Besides, Spectrum Sharing Model is also aresearch hotpoint in Cognitive Radio. By designing appropriate transmission strategy,the spectral efficiency will be significantly improved.At first, the background and significane of this topic have been introduced. Then,the basic theory and previous study of compress sensing and spectrum sharing modelalso have been analyzed. In the last, the contents have been described briefly.In Chapeter Two, the homotopy basedl1norm minimal optimization problem hasbeen introduced and two homotopy based LASSO algorithms have been studied. Theperformance of these two algorithms has been simulated and analyzed both. Besides, aLASSO homotpy based dynamic updating algorithm and a homotopy based reweightedalgorithm have been introduced. Based on their work, two advanced dynamic updatingalgorithms have been proposed according to two dynamic updating models respectivelyand the simulation results show these two algorithms have better performance than theprevious algorithms.In Chapeter Three, sparsity recovery problems of distributed compressive seninghave been studied. In centralized scheme, a Joint-SAMP has been proposed. Based onjoint sparsity structure, this method can get final result by applying parallel computation. Besides, a partial observation matrix based sparsity recovery algorithm also beenproposed. This method can be applied in the situation when the observation matrix inthe fusion center is incomplete. In decentralized situation, at first, a D-LASSOalgorithm has been introduced. Based on this work, a new distributed sparsity recoveryalgorithm has been designed which aiming at recoverying different sparse vectors fromdistributed measurments. Besides, our method has a better recovery performance thanprevious algorithms.In Chapter Four, Spectrum Sharing models have been studied. First, aninterference alignment based spectrum sharing method has been studied and theoptimum power allocation stratey has also been analyzed at same time. Then, aspectrum sharing model which based on dynamic spreading spectrum approach hasbeen introduced and the perforamce also has been analyzed under both interferenceconstraints for PU and quality of service (QoS) constraints for SU. Finally, the impact ofthese constraints on the model has been analyzed too.At last, the contents of this paper have been summarized. The existing bottleneckproblems of compress sensing and spectrum sharing have been introduced briefly.Besides, some suggestions have been proposed for future work.