Resource Allocation Algorithm Based On Beam Hopping For Leo Satellite System

After half a century of development,the terrestrial communication system has already possessed a mature operation mode and achieved large-scale popularization.However,high-quality communication relies on a large number of ground base stations,which is costly for remote areas and is severely affected by disasters.The satellite communication system has the characteristics of wide coverage,and the construction cost has nothing to do with distance.It can well solve the problems of terrestrial communication systems.The satellite communication system has the characteristics of large capacity,low delay,and high spectrum utilization rate,which can well meet people's communication needs.As a typical resource-constrained system,satellite communication system has the limited payload and spectrum resources on the satellite as the key factors restricting its development.Therefore,in order to meet the needs of broadband highspeed services and satellite Internet of things,a multi-beam system is proposed.The multi-beam system adopts multi-beam antenna technology,and its beam has the two major characteristics of spatial isolation and frequency reuse,which is an effective way to solve the problem of resource constraints.In addition,efficient resource management methods are needed to improve system resource utilization,thereby improving system performance.This paper focuses on the problem of resource allocation based on beam hopping in a multi-beam system.The beam hopping has a short working time slot and flexible scheduling,which can better meet the needs of dynamic resource allocation in the communication system.The thesis studies the resource allocation algorithm based on beam hopping in the LEO satellite system.Firstly,a beam-hopping resource allocation algorithm based on multi-objective optimization is proposed.The multi-beam satellite system adopts the beam-hopping mode of work,and uses the multi-objective optimization function in convex optimization to equate the system traffic request to the allocation of time slot resources.According to the characteristics of the inter-relation and mutual restriction of the resources of the frequency,power,and beam dimensions in the system,the overall resource allocation and scheduling are carried out.This paper designs a joint allocation scheme of timeslots,frequencies,and beams based on business priority,fairness,and system capacity maximization constraints.The simulation results show that under different service distribution conditions,the algorithm in this paper is better than the polling beam distribution method in system throughput performance and delay performance.Secondly,this paper proposes a beam-hopping resource allocation algorithm based on deep reinforcement learning.The resource allocation is modeled as a Markov decision process,and the traffic request,waiting delay and beam scheduling in the communication system are abstractly described as state reconstruction.Use the state to construct a neural network,with the goal of optimizing real-time data delay and system throughput,apply the deep Q network algorithm to learn the network,and train a stable beam allocation strategy through a large number of random state inputs and multiple iterations.The simulation results show that the delay and throughput performance of this method is better than random beam allocation algorithm,genetic algorithm-based beam allocation method and the multi-objective optimization algorithm proposed in chapter 3.Moreover,this algorithm has obvious advantages in solving the resource management problems of low-orbit satellite systems with fast service scene changes and complex business types.