Optimization Of Sensing Quality For Cooperative Spectrum Sensing Using Coalition Game Theory

With the rapid development of wireless communication technology, requirementsof spectrum resources and bandwidth by people are more and more urgent; makingthe problem of shortage of spectrum resources is becoming more and more prominent.The cognitive radio technology is just a very effective means to improve the spectrumefficiency，to solve the problem of shortage in spectrum and coexistence amongdifferent kinds of wireless communication networks. Obviously, when cognitive radiospectrum sensing as an important and integral foundation in cognitive radio, to ensureand improve the quality of the spectrum sensing in cognitive radio has become anurgent requirement and one of the hotspots now.According to the basic theory and oversea research status of cognitive radiospectrum sensing and aiming at spectrum sensing quality issues, several keychallenges to improve the quality of the spectrum sensing are summarized, with theconclusions and analyses of various methods to improve the quality of the spectrumsensing as well. On this basis, the study on cooperative spectrum sensing is given,also clarifying its general function modules and working models. With particularemphasis on analyses and researches on cooperative model based on traditionalparallel fusion and coalitional game theory, the model for the cooperative spectrumsensing using coalition formation games is established now.According to the established cooperative sensing model, the perform analysis andthe CF algorithm of cooperative spectrum sensing using coalitional formation gamefor time analysis are given，with explanation and Pseudo code description. Then,some mathematical analyses of the complexity of this algorithm is made, drawing theconclusion that CF algorithm is very complex to implement, especially when thereare plentiful second users in cognitive radio networks. To solve this problem, thestrategy selecting the second users based on their own signal-to-noise ratio (SNR) forcollaborative sensing later is proposed, which can reduce the number of the secondusers for collaborative sensing and eventually optimize the complexity of CFalgorithm. According to the strategy, the Selecting Second-users and CoalitionFormation (SSCF) improved algorithm is proposed too. Then it has been proved by mathematical analyses that compared to the original coalition, the drop value of theutility function for the coalition deleting a second user whose SNR is lower than thethreshold according to the rule for choosing second users for cooperative sensing inSSCF algorithm, is no more than5%. So it is turned out that the sensing quality ofSSCF algorithm is not worse than95%of that of CF.According to some simulation experiments and data analyses for the complexityof the CF algorithm, the paper verifies the CF algorithm is computationally intensiveand result in a lot of time-consuming for sensing and reduce the overall throughput ofthe whole cognitive radio network, especially when there are a larger number ofsecond users in the network. With the corresponding simulations and analyses for thesensing quality of SSCF algorithm, it is demonstrated that the average utility value ofall the nodes in the cognitive radio network by SSCF algorithm, is improvedcompared to that by CF in common; even in the worst case, the former is less than thelater by no more than5%. Thus the conclusion can be drawn that although SSCFalgorithm simplifies the complexity of the CF algorithm by reducing the number ofsecond users participating in the cooperative spectrum sensing, the SSCF algorithmwill not drastically reduce the sensing quality of the CF algorithm.