作　　者： ; ; ;

机构地区： 深圳大学信息工程学院ATR国防科技重点实验室

出　　处： 《信号处理》 2011年第9期1341-1346,共6页

摘　　要： 针对被动传感器网络中的多目标数据关联,提出一种基于D-S证据理论的数据关联新方法。根据被动传感器网络中多目标跟踪的特点,建立适合于被动目标数据关联的6种合成证据(包含速度变化量、航向角变化量、飞行时间变化量、观测可能性、属性变化量与未知),根据各自的自相关函数自适应确定证据权值,通过证据权值与信任函数的乘积动态调整合成中各证据的比重,最后应用多级决策获取对目标观测的关联结果。仿真结果表明,使用自相关函数改进的D-S证据理论的数据关联新方法能够在抑制观测中冲突证据影响的同时,削减关联运算时间,获得被动多目标实时有效的关联结果,性能优于模糊C均值聚类关联方法(FCM)和传统D-S证据理论的数据关联方法。 To associate the information of multi-target in the passive senor networks,a new association method based on Dempster Shafer theory is proposed in this paper.With the characteristics of the passive senor network（irregular time interval and fuzzy description）, this algorithm builds 6 evidence for calculation（including speed changing,Course changing,Time changing,possibility of the observe,attribute member,and Undefined member）.By using the power of different evidence which is calculated from self correlation function,percentage of each member in the association algorithm can be adaptively coordinated by each evidence power and its own belief function.And the association result will be decided by the mulity layer judgement function.The simulation shows that the association improved by the self correlation function based on Dempster Shafer theory in the passive senor networks not only can restrain the obstacle caused by evidence which conflict to other member in evidence group,but also can reduce the calculation time of the association algorithm.It is effective in association with large scale information which is irregular time interval and fuzzy description,and oberversly to prove that this association is better than FCM or traditional Dempster Shafer theory association.

关 键 词： 被动传感器网络 证据理论 自相关函数 数据关联

领　　域： [电子电信] [电子电信]