Research On Optimization Of Satellite Constellation Deployment And Reconfiguration For Rapid Response

Building a large-scale,low-cost,rapid-response low-orbit huge constellation to provide all-day,all-weather,high-performance and time-effective earth information services is the frontier and hot topic in the current space field,representing typical direction in satellite architecture.Traditional satellites constrained by the number,scale and maneuverability,are difficult to meet the coverage requirements of continuity and timeliness,and to provide rapid response to regional conflicts,disaster relief and other emergencies.Considering the development trend of low-cost,high-reliability,rapid deployment and on-demand reconfiguration of future satellite systems,how to optimize the allocation of on-orbit resources and improve the stability,agility and flexibility of satellite constellations has become a frontier issue.Based on the constellation coverage analysis model and multi-objective optimization framework,the dissertation focuses on optimization of satellite constellation deployment and reconfiguration,investigates major optimization issues of rapid launch deployment,large-scale constellation reconfiguration,agile satellite constellation under multi-task constraints,and reconfigurable constellation design under task uncertainty.(1)Constellation rapid launch deployment optimization.In order to build an efficient satellite constellation to cover emergency targets in time and provide rapid response,critical issues such as mission response time,on-orbit response time and other responsiveness indicators should be considered,in addition to coverage performance.In terms of the optimization of asymmetric constellation for zonal emergency observation,an optimization model of constellation rapid launch deployment with multiple responsiveness and coverage indicators is constructed,taking into account the launch window,multi-satellite launch and phasing adjustment.A hierarchical chromosome encoding method that simultaneously features pattern and orbital parameters of the asymmetric constellation is proposed to improve the solution efficiency.In solution analysis and decision making,an interactive Pareto frontier visualization method based on radial axis plots is proposed,which is different from the traditional static view in that it provides an intuitive presentation and analytic comparison of the high-dimensional nondominated solution set.(2)Reconfiguration optimization of large-scale constellation.For a large-scale constellation with hundreds of satellites,the characteristics of reconfiguration tasks are investigated in terms of on-orbit resource allocation and orbit adjustment,and a reconfiguration optimization model for task assignment and on-orbit maneuver is established.The reconfiguration of a large-scale constellation simultaneously optimizes the resource allocation scheme and satellite maneuver variables,which is a complex combinatorial optimization problem with high dimensionality difficult to converge by traditional multi-objective optimization algorithms or hardly meet the computational timeliness requirements.For a quick solution and high optimization efficiency,the adaptive multi-objective evolutionary algorithm based on knowledge-based operators(AMOEA/KO)is proposed to give an efficient search for the high-dimensional decision space by mining the correlation between the design variables and objective functions from the intermediate solution set to form the knowledge-based operators with domain knowledge.In order to prevent the generation of invalid operators and over-intervention of operators,credit assignment and an operator selection strategy based on adaptive pursuit are designed,which can effectively promote the algorithm convergence.(3)Optimization of agile satellite constellation under multi-task constraints.Considering the characteristics of agile satellites,such as multi-objective,multiconstraint and multi-satellite collaboration,the working process of the agile satellite constellation and multiple constraints including task sequence,side-swing imaging,state transition,data transmission link and sequestration are investigated.In order to evaluate the mission effectiveness and optimize the constellation pattern and orbit parameters,an agile satellite constellation optimization framework based on task simulation is proposed to evaluate the mission assignment,side-swing imaging and data transmission processes,and then iteratively optimize the constellation scheme based on the evaluation results.In the framework,such simulation objects as satellites,ground stations and mission centers are constructed to solve the agile satellite task assignment and the evaluation of the constellation overall performance.On the other hand of optimization algorithms,a dynamic evaluation and selection strategy of multiple operators based on dominance is designed,which coordinates multiple operators to avoid the problems of local optimization search or slow convergence by single operator.(4)Reconfigurable constellation design optimization under task uncertainty.To adapt to future needs and build a space-based observation system with sustainable benefits,a flexible design of reconfigurable constellation is required that is different from the "static" design.The uncertainty of emergency missions is analyzed by considering the contingency response requirements during the whole lifetime of the reconfigurable constellation.A multisample Monte Carlo simulation is constructed with an estimated distribution model of historical data to calculate the high confidence overall effectiveness of the reconfigurable constellation.In terms of cost,the total cost of the constellation from development to deployment and from on-orbit to de-orbit is estimated based on the cost estimation relationship,considering payload cost,spacecraft cost,launch cost and related scale effects.On this basis,a multidisciplinary optimization model for reconfigurable constellations is constructed.The optimization framework based on Master-Slaver parallel simulation is used to address the time consuming problem of multisample Monte Carlo simulation.To prevent the optimization search process from stalling,the archival and restart loop mechanism of the ε-NSGA-II algorithm is used to maintain the stability of the algorithm.The simulation results verify that the flexible design of reconfigurable constellations can effectively improve the overall cost-efficiency ratio of constellation construction.In summary,the dissertation focuses on the optimization problem and key technologies for rapid response-oriented satellite constellation deployment and reconfiguration,and proposes the corresponding solution methods.The research work provides a useful exploration of constellation optimization and multi-objective evolutionary algorithms,offering theoretical value for the design,deployment,operation and control,and application of satellite constellations,and laying a solid foundation for further research.