基于离散模糊动态贝叶斯网络的空战态势评估及仿真

Modeling Air Combat Situation Assessment by Using Fuzzy Dynamic Bayesian Network

自动准确地进行空战态势评估是无人作战飞行器(UCAV)自主作战,或有人作战飞行器(MCAV)的辅助作战决策系统必须解决的技术,也是进行威胁评估和战斗决策的基础。为了解决空战态势评估的建模和实现问题,提出了用模糊动态贝叶斯网络实现空战自动态势评估的方法,推导了离散模糊动态贝叶斯网络的推理算法,建立了空战自动态势评估的离散模糊动态贝叶斯网络模型,并进行了仿真验证。仿真结果表明,依据离散模糊动态贝叶斯网络所建立的空战态势评估模型,能够准确地跟踪战场态势的变化,及时发现态势的转换边界,而且在观测值出现偶然误差或者错误时,仍然可以给出正确的评估结果。

Automatic and accurate situation assessment is essential for unmanned combat air vehicles （UCAVs） to conduct and maintain their operations autonomously and effectively. The assessment forms the basis of threat assessment and plays an important role in implementing autonomous control and optimization for UCAVs. It is also a fundamental issue in developing combat decision-making support system for manned combat air vehicles （MCAVs）. A novel approach was proposed as an attempt to tackle this challenging problem. A model based on discrete fuzzy dynamic Bayesian network was derived for UCAVs＇ situation assessment. A detailed theoretical analysis on the model and its inference method was given. Relevant simulation experiments were conducted and the results were discussed. It is shown that the presented model can predict accurately changes of the situation in a varying dynamical environment and has a good performance in terms of effective noise filtering from observations.