Study On The Key Issues Of Cooperative Spectrum Sensing In Cognitive Radio

In recent years, with the rapidly development of wireless communicationindustry and the increasingly clear trend of the diversification, broadband andmultimedia of wireless communication business, the lack of spectrum resources isincreasingly serious, which is a bottleneck to restrict the development of wirelesscommunication in future. Meanwhile, there are three main reasons for this problem:(1) the limited and non-renewable spectrum resources;(2) the traditional fixedspectrum allocation methods led to low spectrum utilization;(3) the growing demandfor spectrum resources. Cognitive radio (CR) is an emerging technology to mitigatethe shortage of spectrum resources to improve the spectrum utilization. By sensing theradio environment, CR system is looking for the un-used spectrum bands licensed byprimary users (PUs) to utilize the idle band by chance and reuse the idle spectrumbands for improving spectral efficiency. Spectrum sensing is not only a keytechnology in CR for protecting the normal communication of primary users withoutinterference, but also the prerequisite of CR system working properly. To improve thepracticality of the spectrum sensing, with the objective of the accuracy and agility forspectrum sensing, a region-oriented spectrum sensing model (RSSM) is established toachieve the extension for the SNR model. Based on RSSM, an improved artificial fishswarm algorithm (AFSA) is proposed as a new attempt for selecting the secondaryusers (SUs).Since the performance of spectrum sensing is seriously influenced by thereceived SNR, the existing studies are based on the assumption of the perfectknowledge of the received SNR. In addition, the value of SNR is determined by thedistance between SU and PU, therefore, it is equivalent to bound the PUs in a specificsmall region. In practices, it is too complex. A reasonable idea is to determine theregion to be perceived according to the known conditions, then, the randomdistribution of the received SINR is obtained by the determined area and the otherknown conditions, and the expectation values of the sensing performance are acquiredby the relevant information of the obtained SNR. Based on the discuss above, a regionoriented spectrum sensing model is set up. Before the modeling, the size of thesensing area is derived based on the principle of fading margin. After that,RSSM-based single user spectrum sensing performance and the RSSM-based decisionfusion collaborative spectrum sensing performance are analyzed, where the effectsfrom the dispersion degree of SUs to the performance of cooperative spectrumsensing are focally analyzed. The study found that under the ‘OR’ criteria, the moredispersed SUs is, the better performance of sensing spectrum is. However, under the ‘AND’ criteria, the more concentrated SU is, the better performance of sensingspectrum is. In addition, regardless of the dispersity of SU, the sensing performancebased on ‘OR’ criteria is better than that of under the ‘AND’ criteria. The simulationresults verified the effectiveness of the proposed approach.Moreover, the selection of SUs is an important part of the cooperative spectrumsensing, which has heavily impact on the sensing performance. Based on theaforementioned discuss, the selection problem of SUs based on ‘OR’ cooperativespectrum sensing is transformed into a nonlinear0-1programming problem with theobjective of maximizing the whole dispersion of SU, which is solved by the improvedAFSA to achieve the selection of SUs. The computer simulations prove that theimproved AFSA is able to quickly achieve the selection of SUs. Compared with theclassical AFSA, the improved algorithm has6times increased optimization successrate, a30%reduction the number of average convergence, a10%increasedtime-consuming. And, the improved AFSA has good robustness for the parameters.Using the improved AFSA about the selection of SUs is an attempt to utilize theswarm intelligence algorithm for the problem of SUs’ selection. Making full use ofthe optimizing speed and robustness of AFSA is the need for spectrum sensing.