一种基于不确定分析多传感器信息动态融合方法.pdfVIP

一种基于不确定分析的多传感器信息动态融合方法1’ 罗本成原魁陈晋龙朱海兵 (中国科学院自动化研究所复杂系统与智能科学实验室北京100080) (E-mail：Bencheng．1uoOmail．ia．ac．cn) 摘要提出了～种基于不确定分析的多传感器动态分布融合方法。首先引入贴近度的概念 对传感器进行动态聚类，接着基于兼容测度实现了组内传感器信息的最优Bayesd邪估计融 合：最后给出了一种基于一致测度的多传感器信息动态融合的方法。通过实验对比分析，证 实了此方法具有较好的实效性和鲁棒性。 关键词不确定分析，格贴近度，动态多传感器信息融合，最优Bayesd棚估计融合． Based Uncertainty AnalysisDynamic DataDistributedFusion Luo YuanKuiChen Zhu Bencheng JinlongHaibing (InstituteofAutomation,ChineseAcademyofSciences,Be#ing100080) (Bmail：Bencheng．1uo@mail．ia．ac．on) AbstractInthis of multi-sensorsdamfusion under papertheproblemmodeling system is the not considersthe intermsof discussed．Modelingproblemonly moasure uncertainty alsoUSeSinformationtomeasurethe and data． quality uncertainty,but compatibilityofsensory in ofthemP．2Stlreofrelatiw terms ofuncertainty,we Takingadvantage proximity firstlyimplement multi-sensors 013 estimation the dynamic Bayesian clustering．Based technology,withconsidering fusion formulti-set．sotsdata in optimalparadigm fusionthesame measul℃ofcompatibility,an group is them1．itlJaI ofsePdor basedonthe impact groups mea乳墉ofconfidence．a presented．Examining novelmodelfor muttGseaxSomfusion isdescribed．Theefficientfusionofdatafrom d、m醐m