Mission Scheduling And Planning Of Multi-target Observations For Spacecraft

Mission scheduling and planning is the key part of autonomous management, which plays an important role in automation of spacecraft. In spacecraft observation, activities of the spacecraft should be planned corresponding to targets that are selected, through the entire process of spacecraft scheduling with the targets. As the number of spacecraft increases and the missions for spacecraft observation becoming more complex, scheduling and planning for spacecraft observation have attracted more attentions of researchers in both academic and engineering domains. Improvements on the research of scheduling and planning for spacecraft observation possess great significance in raising observation profits and reducing running cost for the spacecraft. However, difficulties and challenges arise because of the complicated constraints and the inefficiency of scheduling and planning for spacecraft observation. In this thesis, shortages of current research are summarized, and a feasible mathematical model is established, based on which, the scheduling and planning problem for spacecraft observation under complex constraints are studied systematically, through the sections of temporal constraints analysis, heuristic rapid planning algorithm and optimization for scheduling and planning.A model description for the spacecraft scheduling and planning system is given out based on targets, activities and activity timelines. The logical constraints between targets and the activities of spacecraft are analyzed and proposed, by which, both temporal and resource constraints between targets are transformed into temporal and resource constraints between the activities of spacecraft. The checking methods for every kind of constraints and the specialty for temporal constraints are analyzed.Temporal constraints checking usually occupies large amount of the computing resource, which could greatly influence the efficiency of scheduling and planning. Based on simple temporal network, the characters of temporal constraints existing among the scheduling and planning process are analyzed, both the relationships between two activities and that between an activity and a activity timeline are utilized to form a partial temporal network, upon which a novel temporal constraints checking process is proposed combing with Floyd-Warshall algorithm. The new process could improve the efficiency of temporal constraints checking.An interactive method of spacecraft observation scheduling and planning(SOSP) is proposed, integrating scheduling for targets and the planning for activities, causing continuous planning. For planning, considering numeric character of the constraints, relevancies between activities are defined. A dynamic updating method for relevancies and the mechanism for selecting activities based on relevancies are given out, by which a heuristic planning method are proposed. The method could efficiently reduce planning steps and backup steps, to improve the efficiency of the planning process.Based on SOSP, ant colony algorithm is applied in scheduling and planning optimization for spacecraft observation in this thesis. Analyzing constraints in SOSP, demand forecasting method for different type of resource constraints and dynamic classifying strategy for the targets remain to be selected are given out. Combining the classifying strategy, temporal information and priority constraints, proper heuristic transfer controlling and pheromone updating are designed, and finally a new scheduling and planning optimizing algorithm for spacecraft observation is formed. The new optimization could efficiently increase the profits of observation.At last, to validate the efficiency of the methods and algorithms proposed in this thesis, a simulating platform for scheduling and planning of spacecraft observation is designed and accomplished. The frame of the platform is given out and the components are described as well as their functions. With this platform, the influence on scheduling and planning caused by partial temporal network, the influence on optimization caused by relevancy heuristic, and the results of the optimization are verified.