Research On Routing Algorithm And Topology Control Strategy Of LEO Satellite Networks

With the reduction of satellite manufacturing and launch costs as well as the increase of global communication requirements,Low Earth Orbit(LEO)satellite networks have once again become a research hotspot.A large number of emerging LEO Internet constellation plans have been proposed one after another.The LEO satellite network exists as a supplement to the terrestrial network due to its advantages of wide-area coverage and immunity from natural disasters on the ground.It has become an important means to bridge the "digital divide" and an important part of the future network system.Inter-satellite link is a key technology in the new generation of LEO satellite networks.By establishing intersatellite links between satellites,system capacity can be increased,end-to-end communication delays can be reduced and reliance on ground infrastructure can be reduced.In a LEO satellite network with inter-satellite links,appropriate inter-satellite links are established between satellites through topology control strategies and data transmission paths are constructed through routing algorithms.The design of routing algorithms and topology control strategies directly affects the performance and overhead of the network.This dissertation aims at the research on routing algorithms and topology control strategies of the next-generation LEO satellite networks,while fully considering the characteristics of constellation topology and network services.The main research contents and contributions of this dissertation are as follows:(1)A joint optimization algorithm of flow control and load balancing routing is proposed.Communication between satellites and ground gateways is a typical communication scenario in LEO satellite networks.Due to geographical,political and other factors,the deployment location of gateway stations is usually concentrated,which makes the inter-satellite links and satellite-ground links near the gateway stations easy to become the bottleneck of the network.If only the shortest path-based routing algorithm is used,network congestion will happen.In addition,the traffic that the network can carry is limited by the network capacity.If the traffic entering the network is not constrained,the network congestion will be further aggravated.In addition,the LEO satellite network carries various types of services,thus the network needs to be able to meet the quality of service requirements of different services.Aiming at the scenario where the deployment position of gateway stations is constrained,this dissertation proposes a flow control and routing joint optimization algorithm to achieve network load balancing and quality of service guarantee.The basic idea of the algorithm is to jointly optimize flow control and routing based on the network traffic model and the delay and bandwidth constraints of multiple types of services with the aim of maximizing network throughput.First,the feeder links is incorporated into the network topology and optimized together with the inter-satellite links.A time slice division strategy is proposed to increase the topology connectivity.Then,with the goal of maximizing network throughput,based on the delay and bandwidth requirements of various services,the joint optimization problem of flow control and routing is formulated.Finally,a heuristic algorithm based on simulated annealing is proposed to efficiently solve the joint optimization problem of flow control and routing.The simulation results show that,compared with methods where flow control and routing are optimized independently,the proposed algorithm can improve the network load balancing ability and improve the overall network throughput while meeting the quality of service requirements of various services.(2)A scalable distributed anti-destruction routing algorithm is proposed.Mega constellations can provide higher frequency utilization and become the development trend of the new generation of LEO satellite networks.Among them,the inclined orbit constellations can provide better coverage for densely populated areas and can maintain a relatively stable topology,which is conducive to the design of routing algorithms.However,the inter-satellite links may still fail due to various reasons,therefore,the routing algorithm needs to have survivability.Using flooding-based link state algorithms in mega-constellations will introduce huge computational and signaling overhead,while using local detouring-based distributed anti-destruction mechanisms will increase the extra end-to-end delay.In this dissertation,a scalable distributed anti-destructive routing algorithm is proposed for inclined orbit mega-constellations to achieve a compromise between routing overhead and delay.The basic idea of the algorithm is to reduce the signaling overhead by limiting the propagation range of the link state message and to reduce the path delay by utilizing the regularity of the LEO satellite network topology to detour packets in advance.Firstly,the special topological characteristics of inclined orbit constellation are studied.Utilizing the regularity of constellation topology,multiple primary and alternate paths are calculated for any pair of satellites with low computational cost.On this basis,a link anti-destruction mechanism is proposed.When the link state changes,a restricted flooding mechanism is used to flood the link state in a limited area,which reduces the signaling overhead;after a satellite receives the link state message,the pre-detour mechanism is used to start packet detouring in advance,reducing the extra path delay.In addition,a loop avoidance mechanism is proposed to deal with routing loops with low overhead and a vector-based next-hop selection mechanism is proposed,which comprehensively considers multiple factors to efficiently select the nexthop node.The simulation results show that,for the link failure scenario in megaconstellations,the proposed algorithm reduces the signaling overhead compared with the traditional link state algorithm and reduces the end-to-end delay compared with the local detour-based strategy.It also provides support for load balancing by using its multi-path feature.(3)A delay-optimized inclined orbit constellation topology control strategy is proposed.In inclined orbit constellations,due to the relatively fast relative motion,the inter-satellite links between the two sets of satellites in the intersecting orbits are usually not established,resulting in a long path between the two satellites that are relatively close in space.If such inter-satellite links are established between two sets of satellites,the requirement of tracking and pointing capability of the antenna is high and the overhead of topology switching is introduced.Aiming at the long delay problem in inclined orbit constellations,this dissertation proposes a topology control strategy to optimize network latency performance with low system overhead by comprehensively considering multiple factors of link establishment,including link distance,link angular velocity,link establishment time and topology change frequency.The basic idea of the algorithm is to reasonably select the intersatellite links to provide low-latency paths and extend the time slice as much as possible considering the constraints of the distance,angular velocity and establishment time of the inter-satellite links.Firstly,for a given set of satellites,under the condition that the link distance and angular velocity constraints are satisfied,the links with the longest duration are selected to reduce the frequency of topology change.Then,considering the establishment time constraint,a replacement satellite selection mechanism is proposed to reduce the frequency of topology change by finding the replacement satellites that make the time slice as long as possible.The simulation results show that the algorithm can significantly improve the long-delay problem in inclined orbit constellations while keeping the topology switching frequency and antenna pointing requirements low.