摘要
针对不确定环境下无人机任务决策问题,提出一种基于变结构离散动态贝叶斯网络的自适应推理算法.该算法能够利用软/硬证据和先验信息动态地调整任务决策模型参数,通过推理和参数学习互动的方式使任务决策模型具备适应动态环境的能力.仿真证明,提出的自适应推理算法能够在突发威胁信息不完备、先验参数不精确和先验参数无认知的情况下为无人机任务决策提供保障.
This paper proposes an adaptive inference algorithm for unmanned aerial vehicle mission decision-making under uncertain environment based on structure-variable discrete dynamic Bayesian net- works. The proposed algorithm can dynamically adjust the parameters of mission decision-making model by using soft/hard evidence and a priori information, and make the mission decision-making model suit- able for the dynamic environment by reasoning and interactive parameter learning. The simulation results validate that the proposed algorithm can provide assurance for unmanned aerial vehicle mission decision- making under the conditions of incomplete information of abrupt threats, inaccurate or unawareness of a prior parameters.
出处
《系统工程理论与实践》
EI
CSSCI
CSCD
北大核心
2013年第10期2575-2582,共8页
Systems Engineering-Theory & Practice
基金
国家自然科学基金(61062006)
海南省自然科学基金(612130)
海南大学科研启动(kyqd1209)
关键词
无人机
信息不完备
任务决策
变结构动态贝叶斯网络
unmanned aerial vehicle
information incomplete
mission decision-making
structure-variabledynamic Bayesian networks
