Resource Optimization Of Satellite Network Based On Time-varying Graph

With the increase of satellite earth observation target points,it is of great significance to study how to rationally use the limited spaceborne resources of agile observation satellites to capture more ground information in highly dynamic scenarios.Considering that the small payload of satellites limits the volume of spaceborne equipment and makes it impossible to improve space-borne resources directly from the hardware,satellite network mission planning is an important way to optimize spaceborne resources at present.However,most of the current studies on observation satellites focus on planning observation tasks and downtransmission tasks,ignoring the use of inter-satellite links for data unloading.Timely data unloading is helpful to relieve the onboard memory pressure of observation satellites and improve observation efficiency.Therefore,this paper mainly makes the following two improvements for the scenarios of mission planning of satellite network with limited spaceborne resources: one is to propose an agile observation satellite mission planning strategy in a highly dynamic scenario integrating inter-satellite mission unloading;the other is to build an inter-satellite mission unloading dynamic model combining moving edge computing.In addition,considering that satellite network has the characteristics of high dynamic and predictable topology,time-varying graphs are used to map dynamic scenes into static topology graphs in the above two improvements to simplify the solving process.The specific research content and results in this paper are described as following:(1)Considering that the mission planning strategy of directly transferring observation data from agile observation satellites to ground stations can no longer meet the memory pressure brought by the large increase of observation tasks,the mission planning strategy of unloading observation data to communication satellite constellation is proposed when the onboard memory resources of observation satellites are limited.Firstly,a mathematical model for mission planning was constructed,which combined the periodic charge of spaceborne battery,attitude conversion of agile satellite and time-dependent observation benefits.Then,in order to solve the dynamic problem of agile satellite mission planning,a rolling dual programming framework was established and an improved genetic algorithm based on graph fitness was proposed.The algorithm uses directed acyclic graph,shortest path algorithm and task planning strategy combined with inter-satellite task unload to obtain observation sequence that satisfies both time constraints and spaceborne resource constraints.The simulation results show that the planning strategy considering inter-satellite mission unloading is helpful to obtain the task sequence with the maximum total observation benefit,and the strategy is feasible.The improved genetic algorithm based on graph fitness can effectively shorten the task planning time under the condition of high observation benefit sequence.(2)Considering that the research on scheduling algorithm of agile observation satellite mission planning is difficult to fundamentally solve the problem of limited spaceborne resources,a model is designed in which mobile edge computing server is deployed on the communication satellite constellation to provide computing and storage services for observation satellites.At the same time,in order to solve the problem that the research of moving edge computing is only applied to a single static scene,an unloading strategy is proposed to save the ununloaded data left at the last moment to the next moment,and the correlation between different time slices is established by using the time expansion graph.The simulation results show that the model of inter-satellite mission unloading based on moving edge computing is feasible,and the proposed unloading strategy can effectively improve the unloading rate while ensuring the low cost of observation satellites.