Research On Key Technologies Of Spectrum Map Generation Under Spatial Sparse Samplin

With the continuous development of wireless communication industry,the study of electromagnetic spectrum situation has become the key technology for efficient utilization of electromagnetic spectrum resources and winning the electromagnetic spectrum war.Spectrum map records and stores multi-dimensional data such as time,frequency,location and electromagnetic properties in electromagnetic environment,and visually displays electromagnetic information such as electromagnetic field intensity distribution and spectrum resource distribution.It is applied to frequency reuse,coverage prediction and communication network connectivity,which is helpful to solve complex practical problems in electromagnetic environment and has important research and application value in electromagnetic spectrum situation representation.Aiming at the problem of spectrum map construction under the conditions of sparse electromagnetic spectrum monitoring sensor network and unknown signal source transmission power and location,this paper puts forward a high-precision spectrum map construction algorithm,which provides theoretical basis and technical support for electromagnetic spectrum management and utilization ability.The main work of this paper is as follows:(1)Aiming at the problem of optimal layout of electromagnetic spectrum monitoring sensor network in electromagnetic environment area,this paper proposes a spectrum map construction algorithm based on sensor layout optimization and adaptive kriging.This algorithm uses artificial bee colony algorithm to output the optimized sensor layout without reducing the accuracy of spectrum map construction,and proposes an adaptive kriging algorithm according to the spatial autocorrelation of monitoring data and corresponding geographic information to improve the accuracy of spectrum map construction under the condition of limited sensor resources.(2)Aiming at the influence of noise interference on spectrum map construction in electromagnetic environment,this paper proposes a robust completion algorithm based on lowrank region fusion.The algorithm re-models the electromagnetic spectrum data matrix by introducing the noise factor,and finds the local area lower than the global rank under the constraint of local rank,and completes the matrix,which ensures the robustness of the matrix completion algorithm and realizes the high-precision spectrum map construction of the global area.(3)Aiming at the problem of high-precision construction of electromagnetic spectrum map,this paper proposes an adaptive residual kriging algorithm based on BP neural network.Based on the propagation characteristics of strong correlation of shadow fading,the algorithm uses BP neural network to decompose electromagnetic environment information into path loss components and shadow fading components for estimation,and according to the decorrelation distance of shadow fading components,the monitoring sensor estimation group of unknown points is established through adaptive kriging to improve the overall construction accuracy of spectrum map.