Research On Spectrum Sensing And Spectrum Allocation Algorithm Based On Graph Model Optimization

In recent years,with the widespread application of wireless communication terminal equipment in various fields,the number of terminal equipment has increased dramatically,leading to the depletion of the originally non-renewable radio spectrum resources.Only by improving the utilization efficiency of spectrum resources can the limited spectrum resources be used.To provide more and more network communication services for network equipment.The proposal of cognitive radio technology has opened up new ideas for improving the utilization of spectrum resources.In the entire cognitive radio technology framework,spectrum sensing and spectrum allocation technologies have been widely concerned by many researchers as key technologies to realize the rational utilization of spectrum resources.This thesis proposes a spectrum sensing algorithm and a spectrum allocation algorithm to achieve the purpose of improving the utilization of spectrum resources.The main contents are as follows:Firstly,In recent years,some traditional spectrum sensing algorithms have been widely used in the field of spectrum sensing and have achieved excellent results.However,these algorithms may be greatly affected by noise,the number of cognitive users,and the signal-to-noise ratio.Its detection performance is poor.In order to solve these problems,this thesis proposes a spectrum sensing algorithm based on image K-means clustering analysis.This method first uses the presence or absence of the main user signal to map the two cognitive signal states into an image,and uses image processing to extract the image Feature vector.Secondly,it is sent to the K-means clustering algorithm for training to obtain the classification model.Finally,the trained model is used to classify unknown types of signals to achieve the purpose of spectrum sensing.The simulation results show that the spectrum sensing algorithm based on image K-means clustering analysis proposed in this thesis has better detection performance than traditional spectrum sensing algorithms,especially in environments with low false alarm probability and low signal-to-noise ratio.The next advantage is more prominent.Secondly,In order to solve the problem that the discrete binary particle swarm algorithm is easy to fall into the local optimal solution when solving the spectrum allocation optimization problem,it cannot converge to the global optimal value,and the randomness of the algorithm iterative search becomes stronger and stronger,and the later local search ability is lacking..Based on the coloring model of graph theory,this thesis proposes a spectrum allocation algorithm UMDA-BPSO that combines univariate marginal distribution algorithm and discrete binary particle swarm algorithm.The fusion spectrum allocation algorithm first uses the advantages of the univariate marginal distribution algorithm to have a strong global search capability to pre-optimize the initialized particle swarm to find the approximate range of the optimal solution,and secondly,it uses the binary particle swarm algorithm to perform an accurate search within this range.optimization.Through this fusion method,the limitations of the univariate marginal distribution estimation algorithm and the binary particle swarm algorithm are improved,and the feasibility of the fusion algorithm is proved through simulation,and when the maximum profit of the network is solved,under the same conditions The performance of the algorithm has also been improved accordingly.