Research And Implementation Of Single-star Multi-beam Interference Source Location

With the wide application of satellite communication,the spectrum resources and orbit resources of satellite are becoming more and more strained,and the interference of satellite communication is becoming more and more serious.In order to ensure the safe and stable use of satellite communication,it is necessary to develop the satellite interference source location technology.The antenna carried by Tiantong-1 satellite is a multi-beam antenna,and the single-star multi-beam interference source positioning technology is based on the frequency multiplexing characteristics of the multi-beam antenna carried by the satellite,and there are a lot of gain overlapping areas between adjacent samefrequency beams to achieve positioning.The positioning process does not require special positioning equipment carried by the satellite,and the positioning cost is low.Single star positioning does not need to find auxiliary satellites or carry out parameter synchronization between satellites,which simplifies the positioning process and avoids positioning errors caused by parameter synchronization.In this thesis,based on the theory and simulation of single-star multi-beam interference source positioning technology,in order to improve the accuracy of interference source positioning,the main tasks are as follows:(1)Gain curve fitting of satellite antenna.The positioning equation of single-star multi-beam interference source is based on the gain curve of satellite antenna,and the accuracy of the gain curve directly affects the accuracy of the positioning equation.Because the theoretical gain curve differs greatly from the actual gain curve,the fitting curve of the actual discrete gain points is used to replace the theoretical gain curve when constructing the interference source location equation.This thesis presents four fitting methods for discrete gain points: linear interpolation plus polynomial function fitting,linear interpolation plus Fourier function fitting,cubic spline interpolation plus polynomial function fitting,cubic spline interpolation plus Fourier function fitting.The simulation results show that the fitting curve of the actual gain point is closer to the actual gain curve than the theoretical gain curve.Through theoretical derivation,compared with positioning errors generated by positioning equations constructed by different gain curves under the same gain error,the positioning equation constructed by fitting gain curves proposed in this thesis is less sensitive to gain error,so it is more suitable for positioning interference sources.(2)BP Neural Network(BPNN)solves nonlinear equations.In order to improve the accuracy of the localization equation,BP neural network is used to solve the nonlinear equation.In view of the influence of initial weight and threshold value on the final training effect of BP neural network,Genetic Algorithm(GA)is used to find optimized initial weight and threshold value for BP neural network.When the initial error of BP neural network is low,the training error does not converge.In view of the above problems,it is proposed to adjust the weight and threshold value in time during the training of neural network.The multi-beam interference source location equation was constructed by simulation,and the optimized BP neural network was used to solve the problem,and the error variation chart in the training process of neural network was recorded.By analyzing the error variation chart,it can be verified that the optimized BP neural network can significantly reduce the error of solving nonlinear equations.(3)Validation and evaluation of positioning optimization method.In order to verify the feasibility and effectiveness of the localization optimization method,a visual operation software for interference location was designed.By programming,the gain curve fitting and BP neural network algorithm are integrated into the single-star multi-beam interference source location software.When the positioning software is used,the gain curve can be selected as theoretical gain curve or fitting gain curve,and the algorithm of solving the equation can be selected as particle swarm optimization(PSO)or BP neural network algorithm.Taking Tiantong-1 satellite as an example,input related data parameters into the positioning software.The positioning software selects different gain curves and equation solving algorithms during positioning to verify that the optimization method for locating interference sources can effectively reduce positioning errors.