航天器任务约束处理与规划方法研究

发布时间：2018-04-19 13:11

本文选题：深空探测 + 任务规划　；参考：《北京理工大学》2015年博士论文

【摘要】：在深空探测任务中,由于探测目标距离远、飞行环境不确定、通信延迟等原因,航天器需要解决动态环境中实时控制和观测突发科学目标的问题。提高航天器系统的自主管理能力是解决这些问题的一种有效途径。任务规划技术是实现自主管理的关键技术之一,在航天领域中得到了广泛的应用和发展。采用任务规划技术建立航天器系统的规划问题模型,生成能够自主执行的动作指令序列,可以有效提高航天器完成任务的实时性与可靠性。与此同时,工作环境条件复杂、动作执行受扰动影响等特点也给航天器任务规划问题带来了新的难题和挑战,如何使得规划动作的约束变量满足动态分配的需求成为规划技术研究的目标。本论文针对航天器任务规划中的约束处理过程和规划算法进行了系统的研究和分析,主要研究内容包括以下几个方面:针对航天器任务规划中可动态分配时间约束的处理问题,结合了部分路径一致性最短路径算法和检测冗余边的三角法则,提出了基于部分路径一致性的时间约束计算方法。该方法根据规划动作中的时间约束生成需计算的时间约束网络,并采用部分一致性网络最短路径算法生成的部分路径一致性网络获取顶点间最短距离,简化了计算最短距离的过程。在此基础上,通过分析消去最短距离网络冗余边的过程,提出了采用时间排序的顶点序列过滤部分路径一致性网络的方法,避免了对冗余边的重复检测,可以快速地生成时间网络的最小可分配形式。针对航天器任务规划中资源约束数值的计算问题,结合网络流的增广路方法和预流推进方法,提出了基于时间拓扑排序的资源约束算法。该方法考虑到资源约束网络顶点之间流量分布的特点,利用增广路方法的分层思想将网络中的顶点分为四层。进一步,根据执行时间对所有顶点进行拓扑排序,并采用预流推进方法逐层推进流量,从而计算出资源约束网络的最大流,检验规划结果中资源约束的满足情况。该方法改进了网络最大流的计算过程,有效解决了快速计算规划资源约束数量的问题。同时,结合时间约束网络的可分配形式,提出了检测资源约束是否满足的可分配性条件,提高了规划动作对资源的使用率。针对航天器对执行动作时动态分配约束变量的需求,结合了时序回溯任务规划算法,提出了基于时间约束可分配度的启发式规划方法。该方法首先通过分析时序回溯规划算法的计算过程,确定了在规划空间扩展过程中时间约束网络的变化,并提出了时间约束可分配度的概念。然后,建立了时间约束可分配度的规划启发式函数,给出了以时间约束可分配度作为性能指标的规划过程。结合本文研究内容,根据对航天器任务规划方法进行仿真验证的需求,给出了一套航天器任务规划验证系统的设计方案。首先给出了规划验证系统的软硬件结构和各模块功能,建立了各模块间的数据流程和工作流程。在此基础上,结合本文提出的规划系统实现了任务规划模块,并针对航天器实际工作情况实现了航天器自主在轨运行模块和自主管理模块。最后,建立了描述航天器任务的规划模型,并通过仿真实验验证了航天器系统实现约束处理和任务规划的能力。
[Abstract]:In the deep space detection task, due to the distance of the detection target, the uncertainty of the flight environment and the delay of communication, the spacecraft needs to solve the problem of real-time control and observation of the unexpected scientific targets in the dynamic environment. It is an effective way to improve the autonomous management ability of the spacecraft system. One of the key technologies of master management has been widely used and developed in the space field. The task planning technology is used to establish the planning problem model of the spacecraft system, and to generate the action instruction sequence that can be executed independently, and it can effectively improve the real-time performance and reliability of the spacecraft to complete the task. At the same time, the working environment is complex, The characteristics of action execution affected by disturbance also bring new challenges and challenges to the task planning of spacecraft. How to make the constraint variables of the planning action meet the dynamic allocation requirement has become the goal of the research of the planning technology. This paper systematically studies the constraint process and the planning algorithm in the mission planning of the spacecraft. The main research contents include the following aspects: Aiming at the problem of dynamic allocation of time constraints in the mission planning of spacecraft, combining the shortest path algorithm of the partial path and the triangle rule of redundant edge detection, a time constraint calculation method based on the consistency of the partial path is proposed. The time constraint in the time constraint network is generated, and the shortest distance between the vertices is obtained by the partial path consistency network generated by the shortest path algorithm of the partial consistency network, and the shortest distance is simplified. On this basis, the time scheduling is proposed by analyzing the process of eliminating the most short distance from the redundant edge of the network. The method of sequence vertex sequence filtering partial path conformance network avoids repeated detection of redundant edges and can quickly generate the minimum distributable form of time network. In view of the problem of computing resource constraints in spacecraft mission planning, combined with the augmented approach and preflow method of network flow, a time extension based on time extension is proposed. This method takes into account the characteristics of the flow distribution between the vertices of the resource constrained network, and divides the vertices in the network into four layers by using the hierarchical idea of the augmented road method. The maximum flow of the source constraint network tests the satisfaction of resource constraints in the planning results. This method improves the computing process of the maximum flow of the network, and effectively solves the problem of the number of constraints for fast computing planning resources. At the same time, the distributable condition of whether the resource constraints are satisfied is proposed by combining the distributable form of the time constrained network. In view of the requirement for the dynamic allocation of constraint variables in the execution action of the spacecraft, a heuristic programming method based on time constraint allocation is proposed in view of the requirement of the dynamic allocation of constraint variables for the execution action of the spacecraft and the time sequence backtracking task planning algorithm. This method first analyzes the calculation process of the time series backtracking planning algorithm and determines the planning. In the process of space extension, the time constraint network is changed, and the concept of time constraint distributive degree is proposed. Then, the planning heuristic function of time constraint distributable degree is established, and the planning process of the time constraint distribution is given as the performance index. The design of a set of spacecraft mission planning verification system is given. First, the software and hardware structure of the planning verification system and the functions of each module are given, and the data flow and work flow between each module are established. On this basis, the task planning module is realized with the system proposed in this paper. The actual operating conditions of the spacecraft have realized the autonomous operating module and independent management module of the spacecraft. Finally, a planning model describing the mission of the spacecraft is set up, and the ability of the spacecraft system to realize the constraint processing and task planning is verified by the simulation experiment.

【学位授予单位】：北京理工大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：V47;V448.2

【相似文献】