摘要： 为解决无人机(UAV)在突发威胁下的环境感知问题，提出一个基于后验概率支持向量机(PPSVM)的时变贝叶斯网络(TVBN)态势评估模型，利用PPSVM完成突发威胁观测信息的分类处理，将分类信息的不确定度作为态势评估的证据，使用概率推理实现动态环境下的态势评估。以UAV规避空中突发威胁为仿真背景验证该模型的正确性，并表明其评估结果能真实地反映环境变化情况。

关键词: 无人机, 态势评估, 时变贝叶斯网络, 后验概率支持向量机, 突发威胁

Abstract: In order to solve the problem of situation assessment for Uninhabited Aerial Vehicle(UAV) in emergent threats environment, this paper presents a situation assessment model based on Time-varying Bayesian Networks(TVBN). The Posteriori Probability Support Vector Machine (PPSVM) is used to make the uncertain classification of pop-up threat information, and these classification results are used as input information into the assessment model as evidences. Probabilistic reasoning is proposed to complete the situation assessment in dynamic environment. The rationality of the model and the accuracy of situation assessment are verified by simulating the task that the UAV detects the emergent threats.

Key words: Unmanned Aerial Vehicle(UAV), situation assessment, Time-varying Bayesian Networks(TVBN), Posteriori Probability Support Vector Machine(PPSVM), emergent threats

中图分类号: 

• TP181