量测不确定下多传感器一致性数据融合算法  

Multi-sensor consistency data fusion algorithm in observation uncertainty

作　　者：付春玲[1] 宫德龙[1] 李捷[2] 

机构地区：[1]河南大学基础实验教学中心,河南开封475004 [2]河南大学计算机与信息工程学院,河南开封475004

出　　处：《传感器与微系统》2013年第7期113-115,119,共4页Transducer and Microsystem Technologies

基　　金：国家自然科学基金资助项目(60972119;61170243);河南省创新人才培养计划资助项目(114100510001);河南省高等学校青年骨干教师计划资助项目(2010GGJS-041)

摘　　要：针对量测不确定下多传感器信息采集系统中数据有效融合处理,提出一种量测数据一致性融合算法。首先,引入马氏距离实现两传感器量测数据一致性的近似估计;然后,结合模糊理论中隶属度函数的性质定义一种度量局部量测信息提取和利用效率的支持度函数,进而构建综合全局量测数据间相互支持程度的支持度矩阵;最后,通过对支持度矩阵的计算求解各传感器量测数据在评估系统状态估计中的权重,并通过加权融合方式实现对系统状态真值的估计。理论分析和仿真实验验证了算法的有效性。Aiming at data effective fusion process of multi-sensor information gathering system in observation uncertainty ,a multi-sensor consistency data fusion algorithm is proposed. Mahalanobis distance is introduced to approximately estimate the consistency degree between two sensor observations. Combining with the properties of membership function in fuzzy theory, a new support degree function is defined to extract and utilize the local observation information, and the support degree matrix which can synthesize global observation datas is construction. By calculating support degree matrix to solve the weight of each sensor observation data in evaluation system state estimation to estimate the real value of system state by weight fusion. Theoretical analysis and simulation ext3erimental results verify the efficiencv of algorithm.

关 键 词：多源信息融合 量测不确定 冗余互补信息 一致性融合 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]