Research On Dynamic Allocation Method Of Beam Hopping Resources In GEO Satellite Communication System

In recent years,with the development of Space-Earth Integrated Network and Internet over Satellite,people’s traffic demand for satellite communication has increased explosively.High throughput satellites adopt multipoint beam and frequency multiplexing technology,which makes them have greater communication capacity than traditional multi beam satellites and can better meet the current growing traffic demand.However,in high-throughput satellite communication systems,the power and frequency resources allocated to each beam are relatively fixed;In addition,due to the spatial-temporal heterogeneity of satellite traffic distribution,the diversity of traffic types and other factors,the above problems make the satellite resource allocation inflexible,resulting in a waste of communication resources.Beam hopping technology can allocate resources in the four dimensions of power,frequency,space and time,and the resource scheduling is more flexible.It provides a new idea for the flexible and dynamic allocation of satellite resources,and is the key technology of the future high-throughput satellite communication system.However,the existing resource allocation algorithm in the beam hopping satellite communication system is not suitable for the scenario of real-time matching the dynamic changes of traffic.This paper mainly studies the inter cluster power optimization and resource dynamic allocation in GEO beam hopping satellite communication system,and completes the design and implementation of the beam hopping resource dynamic management and access simulation software.Firstly,this paper summarizes the basic architecture of the beam hopping satellite communication system,satellite channel characteristics and the beam hopping time slot model.On this basis,a beam hopping forward link traffic model is established,and a variety of power resource optimization algorithms are simulated with the power resource allocation as the optimization objective.Simulation results show that the power optimization algorithm can improve the utilization of power resources among clusters.Then,in the beam hopping forward link traffic model,a beam hopping resource allocation algorithm based on deep reinforcement learning is proposed to minimize the packet transmission delay.The algorithm models the satellite resource allocation module and the forward link packet traffic scenario as agents and environments respectively,and designs a set of beam hopping patterns based on interference avoidance criteria as the action set of agents.The decision-making neural network of the agent is obtained through the continuous interactive training between the agent and the environment,so as to solve the optimal resource allocation scheme of the beam hopping satellite system.Simulation results show that,compared with the traditional allocation algorithm,this method can effectively reduce the average packet transmission delay and improve the system throughput.Finally,in order to demonstrate and verify the resource management of beam hopping satellite communication system in complex environment,the design and implementation of beam hopping resource dynamic management and access simulation software are developed based on MATLAB platform.The complex environmental factors such as link rain attenuation,spectrum resources,ground wave level interference and user terminal beam switching are considered in the scene design of the simulation software.In addition,considering the user access management in the scene,an access management strategy based on user priority is proposed,and the resource allocation is realized through the beam hopping resource management module.The simulation test shows that the simulation software can realize the functions of resource allocation,user access management and user terminal beam switching,and can effectively verify the beam hopping resource management method.