Research On The Technologies Of Compressed-sensing-based Wideband Spectrum Sensing And Resource Allocation In Cognitive Radio Networks

The rapidity and reliability of spectrum sensing in cognitive radio (CR) directly affect itspractical application. In CR spectrum sensing, traditional narrowband spectrum sensing cannot meetthe demand of wireless spectrum, and wideband spectrum sensing has become the research hot spotin CR. However, wideband spectrum sensing based on Nyquist sampling theorem puts forward agreat challenge to data sampling rate, data storage and data transmission cost. The emergence ofCompressed Sensing (CS) theory breaks up the limitation of sampling rate in traditional Nyquistsampling theorem, which makes wideband spectrum sensing technique a great application prospect.CS-based multiuser cooperative compressive wideband spectrum sensing schemes in cognitiveradio networks (CRN) are mainly studied in this dissertation. On this basis of it, multiple resourceallocation problems for the utilization of multiple secondary user (SU) detected spectrum holes areinvestigated. That is, the joint optimization of multi-user subcarriers (subchannels) and powersallocation. Specific works are shown as follows.The research background and the significance are introduced in the chapter1. In this chapter,cognitive radio network (CRN) and its key technologies are briefly introduced. Moreover, theresearch progress of CS theory and CR wideband spectrum sensing based on CS are also illustrated.Meanwhile, this paper presents the current research of CRN resource allocation.The basic principles of CS theory are introduced in chapter2. It mainly focuses on threeprocedures in CS theory, which includes signal sparse transformation, measurement matrix designand signal reconstruction algorithm design. On the basis of it, the dissertation illustrates three jointsparse models (JSM-1, JSM–2and JSM–3) in distributed compressed sensing (DCS) with theirapplication scenarios in detail.Wideband spectrum sensing based on CS in CRN is studied in chapter3. The dissertationdiscusses the traditional spectrum sensing algorithm in single cognitive user and multiple cognitiveuser scenarios respectively. On the basis of it, a fast multiuser collaborative compressive spectrumsensing algorithm based on blind sparsity level matching is proposed in this dissertationm, whichcombines CS and spectrum detection. The algorithm greatly reduces the computational complexityand reduces the wideband spectrum sensing time with the premises of detection performancemaintenance. The algorithm meets the target of real-time requirement and reliability in CRN, whichprovides a certain reference for the further research in CR wideband spectrum sensing based on CS.Multiuser multi-resource allocation methods based on cognitive orthogonal frequency divisionmultiplexing (OFDM) are studied in chapter4. Multiuser multi-resource allocation models based onrate adaptation (RA) criterion and margin adaptation (MA) criterion are introduced respectively. On the basis of recent researches, a joint subcarrier and power allocation algorithm consideringsecondary user (SU) rate fairness is proposed, which is based on RA criterion. The algorithm isseparated into two steps, with subcarrier and power allocation respectively. The proposed schemecan maximize cognitive link capacity with the consideration of the fairness between all SUs.Simulation results are given with analysis. Meanwhile, initial centralized CRN is firstly constructed,and the distributed multi-hop system model is constructed when new SUs who preempt the optimalchannel for communication enter into the network. The dissertation also analyzes the systemperformance in detail. Numerical results are given in accordance with the rationality of systemmodel construction.Finally, the conclusion and further work prospect are given in chapter5. This dissertationexplores multiuser multi-resource allocation based on CS theory. The research results providepositive meaning for the further researches.