Enhancing Context-Aware Reactive Planning for Unexpected Situations of On-Orbit Spacecraft
Bingqing Shen, Li Da Xu, Hongming Cai, Yu Han, Pan Hu, Lihong Jiang, Jingzhi Guo
Abstract
Spacecraft software is a complicated software system integrating many subsystems of different disciplines. Supporting on-orbit missions needs the knowledge of industrial information integration. To ensure mission success, reactive planning is a critical function in solving dynamic problems during task execution, relying on the knowledge of situation for optimal decision-making. However, it may unable to correctly identify the operational context in unexpected situations. This article solves the unexpected situation problem for reactive planning. It proposes a context-awareness model enhancement framework to identify the triggers of critical situations based on event evolution-based analysis. The framework includes the process of unexpected situation discovery, context identification, and context refinement. With this approach, preventive operational resilience can be achieved. Moreover, a symbiotic computing paradigm and a flexible inference engine are devised for addressing the on-orbit spacecraft-specific challenges. Also, an agent-based reactive system design and a system integration architecture is provided for implementing the proposed approach from both autonomy and information integration perspective. Experiment results show that the proposed approach is effective and efficient to solve the unexpected situation issue. This article offers a crucial insight to context-aware decision-support in space applications.