变量可信性策略和概率准则在传感器融合的应用  

Probability Criterion and Quantity Creditability Tactics for Sensor Fusion

作　　者：史旭光[1] 孙金玮[1] 

机构地区：[1]哈尔滨工业大学自动化测试与控制系,哈尔滨150001

出　　处：《仪器仪表学报》2002年第z2期729-730,共2页Chinese Journal of Scientific Instrument

摘　　要：提出了一种基于变量可信性策略的多层信息融合方法,这种算法采用多层信息融合结构,具有周期性和递归性的显著特点。这种新的计算方法采用矩阵变换,从而方便了程序编制工作。还试图建立变量可信性策略和概率标准之间的内在联系,在此基础上,对这两种方法在效率,局限性及符号标注方面做出评估。This paper proposes a multi-layer fusion method, which might be entitled as quantity creditability tactics. The multi-layer framework of the algorithm ensures its calculation of many merits ,such as recursive or periodical fusion, high-lighting some particular quantities,etc. Compared with probability criterion, the new method employs the calculation with matrix conversions, thus is more convenient for computer to program. More ever, the research attempts to setup the inherent relationships between the two methods in order to make clearer comments on their properties in terms of efficiency, limitations and notation.

关 键 词：数据可信性 模糊算法 多层信息融合 

分 类 号：TP2[自动化与计算机技术—检测技术与自动化装置]