Research On Spectrum Sensing And Spectrum Allocation Based On Swarm Intelligence Optimization Algorithm

The traditional spectrum sensing method and spectrum allocation technology cannot meet the requirements of spectrum utilization in the current complex communication environment.In recent years,spectrum sensing technology and dynamic spectrum allocation technology based on machine learning have developed rapidly.It has incomparable advantages over traditional spectrum sensing and spectrum allocation algorithms,and effectively improves spectrum sensing efficiency and network benefits during allocation.However,there are problems of low spectrum detection accuracy and poor spectrum allocation under low SNR.This dissertation takes cognitive radio technology as the research background.Firstly,the defects and shortcomings of traditional spectrum detection algorithms are compared and analyzed,and the allocation model of spectrum allocation strategy is discussed.According to the actual needs,support vector machine(SVM)is selected for spectrum sensing.Then,the problem of kernel space parameter accuracy setting in SVM spectrum allocation is studied in depth,and a single-user spectrum sensing algorithm based on kernel space optimization SVM is proposed.Finally,the improved Archimedes optimization algorithm and grey wolf algorithm are used to study spectrum allocation.The experimental simulation shows that the design requirements are met.The research work of this thesis is mainly reflected in the following three aspects:(1)Aiming at the problem that the traditional spectrum sensing algorithm is susceptible to the number of cognitive users and noise fluctuations,resulting in poor performance of spectrum sensing detection,a single-user spectrum sensing algorithm based on kernel space optimization SVM is proposed.The kernel space optimization,SVM and spectrum sensing are combined to denoise the signal and extract its features.The feature vectors are constructed for training and learning,and the improved tunicate swarm algorithm is used to optimize the training search to obtain the best kernel function parameters σ and penalty coefficients C.Simulation results show that the algorithm can effectively improve the detection rate at low SNR and has good robustness.(2)Aiming at the defects of Archimedes optimization algorithm in spectrum allocation,such as slow optimization speed,low precision and easy to be affected by local extreme points,a spectrum allocation scheme based on Circle Archimedes optimization algorithm is proposed based on graph theory model.Firstly,the initialization process of the algorithm is improved by Circle chaotic disturbance.Secondly,the levy flight strategy is used to optimize the global search ability.Finally,the piecewise weight strategy is introduced to optimize the position update process.The acceleration of the immersed individual is discretized by the Sigmod function and applied to spectrum allocation.The simulation results show that the spectrum allocation strategy has a great improvement in solving the feasibility of spectrum allocation and solving network benefits.(3)Aiming at the problem that the gray wolf algorithm has strong global search ability in solving the spectrum allocation problem but has irregular initial solutions and poor local search ability,a spectrum allocation strategy based on the gray wolf optimization algorithm is proposed based on the graph theory model.Firstly,the strategy introduces Sin chaotic initialization and reverse learning strategy to redefine the initialization of gray wolf population.Secondly,the random walk strategy is introduced to optimize the local search.Finally,the Sigmod function is used to binarize the gray wolf position update and apply it to spectrum allocation.The simulation shows that the above spectrum allocation based on the optimized grey wolf algorithm has faster convergence speed and stronger optimization ability than the classical swarm intelligence algorithm spectrum allocation,which can effectively improve the network efficiency of spectrum allocation.