【摘要】 未认知与不确定性是深空探测任务的基本特征.本文基于战略导向的体系化基础研究,建立了一种面向科学价值最大化的探测场景和以可靠性为核心技术基础的深空探测自主运行可信性技术体系.分析研究了深空探测场景下的可靠性概念;面向精确感知、最优计算、准确决策、快精准执行的目标要素,提出了深空探测自主运行的可信性体系框架以及“需求-认知-工程”总体技术架构;针对自主运行可信性的关键技术难点,开展了可靠性导向的多物理场、强耦合白盒建模,复杂网络故障传播机制分析, COTS元器件深空探测应用的高可靠保证,以及“模型+数据+知识”一体的融合机制分析等研究.对该技术体系的关键技术验证策略及其最小系统在卫星星座中的应用进行了验证,结果表明,所提出的技术体系具有较高的工程价值.更多还原

【Abstract】 The “unrecognized and uncertain” is the basic feature of deep space exploration missions. A technology system of dependability for autonomous operation in deep space exploration was proposed based on the principle of “scientific value maximization”-oriented exploration and reliability as the core technology. The concept of reliability in the scenario of deep space exploration was analyzed and involved. A dependability systematic framework for autonomous operation of deep space exploration and its overall technical architecture of “demand-cognition-engineering” were proposed, which include precise sensing, optimal calculation, accurate decision-making and “quickly, precisely, exactly” execution.The reliability-oriented multi-physics deeply coupled white-box modeling and its high-fidelity technology,the complex network fault propagation mechanism and its high-confidence technology, the high-reliability assurance route for the application of COTS components in deep space exploration, the fusion mechanism of “model + data + knowledge” and its dynamic evolution technology for multi-agent health state management based on multiple fetal sensing phenomena were studied. The key technology of this technical system and the application verification of its minimum system in the satellite constellation were introduced, and the results show that the technology system proposed has high engineering value.更多还原