基于动态贝叶斯网络的战斗目标综合推理识别  

Fusion Inference Identification of Combat Targets Based on Dynamic Bayesian Network

作　　者：张伟豪[1] 陈怀新[1] 崔莹[1] 王连亮[1] 

机构地区：[1]中国西南电子技术研究所,成都610036

出　　处：《电讯技术》2012年第6期893-897,共5页Telecommunication Engineering

摘　　要：针对空战战场环境下的目标可靠识别,提出了基于动态贝叶斯网络的战斗目标综合推理识别方法。分析了目标属性的多层次及状态变量关系,提出了层次化的战斗识别动态贝叶斯网络拓扑结构及其参数设定方法,并运用时间片联合树算法进行不确定性动态推理,实现动态的目标属性判断与识别。仿真结果给出了目标的多层次属性信息,验证了模型的有效性。To realize reliability identification of targets in air battlefield, integrated inference for combat target identification is proposed based on dynamic Bayesian network（DBN）. Combat targets＇ attribute variable and its states are analysed, a hierarchy DBN is built for combat identification, then the way of obtaining network param- eters is discussed. Uncertain dynamic reasoning is made based on Joint Tree algorithm of time slice and decision of targets＇ attribute is made dynamically. Simulation study gives targets＇ multi- hierarchy attribute and shows that the proposed method is effective.

关 键 词：目标综合识别 动态贝叶斯网络 不确定推理 时间片联合树算法 

分 类 号：TN97[电子电信—信号与信息处理] TP181[电子电信—信息与通信工程]