Research On Spectrum Sharing Algorithms In Unmanned Aerial Vehicle Network

Unmanned aerial vehicle network faces the challenges of constantly changing electromagnetic environment and low efficiency of static spectrum allocation.Spectrum sharing methods can realize real-time perception of the electromagnetic environment and dynamic management of spectrum resources.It is an effective means to solve the above problems up to now.Spectrum sensing and spectrum access are two of key technologies for realizing spectrum sharing.This thesis focuses on the specific approach of applying spectrum sharing to UAV network,studies the low accuracy of spectrum sensing and low spectrum access efficiency problem when the sharing occurs between flying ad-hoc network and ground radio terminal network.For the problem of low accuracy of spectrum sensing,research is carried out from two aspects: the distribution of perception revenue and the fusion of multi-source perception data.In addition,the dynamic spectrum access algorithm supported by spectrum sensing data is studied.The details are as follows.(1)Research on the spectrum perception revenue distribution algorithm in UAV networking.The thesis builds a blockchain based cognitive UAVs spectrum sensing model,analyzes the relationship between the reputation value and the number of member UAVs,gives the criteria for member UAVs to participate in the sensing task and proposes a pricing method based on reputation value and a revenue distribution algorithm,which overcomes the problem that contribution differences are not considered in the revenue allocation.Make the sensing revenue proportional to the reputation value.The simulation results show that the proposed algorithm can effectively enhance the sensing revenue of member UAVs with high reputation and is conducive to improving the accuracy of spectrum sensing.(2)Research on the multi-source perception data fusion algorithm in UAV network.The thesis constructs a quotation model based on member UAVs’ perceived location and reputation value,proposes a multi-source perception data fusion algorithm based on reputation value quotation,which overcomes the lack of consideration of data quality differences and perceived data discreteness in the spectrum data fusion algorithm.The simulation results show that the convergence speed and accuracy of the proposed algorithm are better than the data fusion algorithm based on the principle of large numbers.(3)Research on the spectrum access algorithm in UAV networking.In this thesis,a spectrum access model based on SNR restriction is constructed,and a prior knowledge-based collaborative learning mechanism and a multi-leader UAV collaborative spectrum access algorithm based on this mechanism are proposed,which overcomes the problems of excessive iterations of the spectrum access algorithm and easily falling into local optimum.The simulation results show that the convergence speed and average throughput of the proposed algorithm are better than those of the spectrum access algorithm based on the common learning mechanism.