Abstract:

It is difficult to predict the combat intention of an air-defense target. Hence， an intention recognition method based on the dynamic Bayesian network and template matching is proposed to predict the combat intention of air-defense targets. Specifically， target feature information based on real-time perception， domain expert knowledge， and accumulated historical combat situation data are used for the reasoning of current combat operations of an air-defense target by the dynamic Bayesian network. Then， a possible sequence of operations is predicted according to the current combat operations of the target. On this basis， the combat intention of the target is predicted by the situation template matching method. The simulation shows that the proposed method can effectively predict the intention of the air-defense target and obtain a higher recognition rate.

Key words: air-defense target, intention recognition, dynamic Bayesian network, template matching

摘要：

对空中目标的作战意图进行预测是目标意图识别中的难点，提出了一种基于动态贝叶斯网络和模板匹配的意图识别方法，实现空中目标作战意图的实时预测。首先，综合利用实时感知的目标特征信息、领域专家知识以及积累的历史作战态势数据，并采用动态贝叶斯网络推理来袭目标当前的作战行动；然后，根据目标当前的作战行动预测目标可能的行动序列；进一步基于目标可能的行动序列，采用态势模板匹配的方法，实现对目标作战意图的实时预测。仿真结果表明，所提方法能够有效解决意图识别中的意图预测问题，进一步提升了意图识别的准确率。

关键词: 空中目标, 意图识别, 动态贝叶斯网络, 模板匹配