Research Of Constraint Processing And Planning Method In Spacecraft Mission

In deep space exploration mission, due to the long distance of exploration targets, uncertainties of flight environment, communication delays and other factors, the spacecraft needs to solve problems of real-time control in dynamic environment and observing unexpected scientific objectives. Mission planning is a key technology to achieve autonomous management, which has been widely used and developed in aerospace field. Mission planning techniques should be used to describe planning model of spacecraft and generate activities instructions sequence which can be autonomously executed, so that, the spacecraft can improve the real-time and reliability when it is completing mission. At the same time, the complex conditions of working environment and dynamic changes of activities executing results also bring new problems and challenges to spacecraft mission planning.This dissertation deeply studies constraint processing method and planning algorithm in spacecraft mission planning. The research works in this dissertation are as follows:For the problem of satisfying dynamic allocation demand of activities time in spacecraft mission planning result, combining with partial path consistency shortest path algorithm and triangle rule which used to generate minimal dispatchable temporal constraint network, the temporal constraint computing method based on partial path consistency is proposed. This method generates temporal constraint network to be calculated according to temporal constraints in planning activities, and uses partial path consistent metwork generated by partial path consistency shortest path algorithm to obtttain the shortest distance between vertexes, which is instead of All-Pairs-Shortest-Path algorithm calculating distance of all the vertexes, so that the process of obtaining shortest distance is simplified. On this basis, according to filtering redundant edges of distance graph based on triangle rule, the method of filtering partial path consistency network using time sorting vertex sequence is proposed, which avoids duplication detection of dominated edges. The dispatchable form of temporal network can be quickly generated.For the problem of calculating resource constraint value in spacecraft planning, combing with augment path method and preflow push method in network max flow theory, the resource constraint method based on time topological sorting is proposed. This method generated resource constraint network to be calculated according to resource constraint needed to be satisfied in planning. Considering the characteristics of flows between resource constraint network vertexes, the vertexes in network is divided in to four layers according to augment path method. Further, all vertexes is topological sorted by execution time, and the flow is pushed according to serial number of vertex with preflow push method. So that, the maximum flow of resource constraint network is calculated, and whether resource constraint is satisified is checked. This method improves the general algorithm to calculate network maximum flow, and gives an effective solution of calculating planning resource constraint value. Meanwhild, combined with dispatchable form of temporal constraint network, the method of dynamically allocating resource variable is given.For the demand of dynamically determining constraint variable by spacecraft in activities execution, combined with chronological backtracking planning algorithm, a heuristic planning method based on constraint dispatchably is proposed. Firstly, this method analyses calculating process of chronological backtracking planning algorithm, and determine the changements of temporal constraint network in expansion of planning space. The concept of temporal constraint dispatcable value is proposed. Then, according to the searching process of planning space, the heuristic function based on constraint dispatchably is established. The planning process which uses temporal constraint dispatchable value as performance indicators is given. In addition, for the problem of processing different types of constraints in planning process, the constraint satisfy problem module, temporal constraint processing module, resource constraint processing module, and the propsed heuristic planning method is conjuncted to form a complet planning system.Combined with this paper, according to the demand of simulation and verifying spacecraft mission planning method, the design scheme of spacecraft mission planning and verification system is given. Firstly, the hardware and sofrward structure and functions of each module of planning and verification system is given, meanwhile, the dataflow and workflow between modules is established. On this basis, combing with the proposed planning system, the mission planning module is achieved. The autonomous operation module and autonomous management module is also achieved by actual working conditions of spacecraft. Finally, planning model of a spacecraft problem is established, and solved by simulation experiments. The ability of spacecraft system to achieve autonomous mission planning is verified.