山地条件下的无人机气象威胁度建模与评估方法  

A Method of Weather Threat Level Modeling and Assessment for UAVs Mountainous Region

作　　者：王俊[1] 周树道[1] 叶松[1] 朱国涛[1] 程龙[1] 

机构地区：[1]解放军理工大学气象学院,南京211101

出　　处：《电光与控制》2012年第5期108-112,共5页Electronics Optics & Control

基　　金：国家自然科学基金(40976062)

摘　　要：针对无人机在山地条件下飞行时气象威胁因素复杂多变,提出一种基于贝叶斯网络的山地气象威胁度建模与评估方法。结合专家知识与EM算法构建贝叶斯网络模型,利用多树传播推理算法对山地气象要素进行威胁等级评估。评估结果表明,基于贝叶斯网络的气象威胁度评估能够正确直观地划分山地气象要素对无人机的威胁等级,对无人机智能决策具有重要意义。In view of the complicated and changeable weather threat when the UAV flying in mountainous region,we proposed an algorithm for mountainous region meteorology threat level modeling and assessment.A Bayesian Network model was constructed combining with the expert knowledge and EM algorithm,and the Polytree Propagation was used for mountainous region meteorology threat level evaluation.Simulation results showed that the algorithm can correctly rank the mountainous meteorology threat level to UAV,which is important for intelligent decision-making of the UAV.

关 键 词：无人机 山地气象 贝叶斯网络 EM算法 多树传播算法 

分 类 号：V279[航空宇航科学与技术—飞行器设计] TP391.9[自动化与计算机技术—计算机应用技术]