模糊传感器与区间型多属性决策的信息融合方法  

Method of Information Fusion Based on Fuzzy Sensor and Interval Multi-Attribute Decision-Making

作　　者：吴瑕[1] 周焰[1] 

机构地区：[1]空军雷达学院预警监视情报系,武汉430019

出　　处：《宇航学报》2011年第6期1409-1415,共7页Journal of Astronautics

基　　金：国家自然科学基金项目资助(40101019);空军雷达学院科研基金项目资助(LYYY200807)

摘　　要：对于日益复杂的信息环境和变化的目标特性,现有的多传感器信息融合方法大多为"静态的",较少考虑传感器可信度变化以及测量过程中的多维特征指标权重的时效性给融合结果带来的影响。为了解决此问题,提出了一种基于模糊理论与区间型多属性决策的信息融合方法。该方法从各传感器对模糊命题支持度的一致性来定义其信息质量优化度,用区间数与多属性决策理论来定义特征识别综合置信度,并从这两方面客观确定了传感器的融合权重,较好地解决了对传感器信息与目标特征权重为模糊与不确定时的合理评价,以此构建了一个智能优化决策层融合识别模型。最后以弹道目标识别实验证明了此方法有较好的时效性和准确性。For the complex information environmet and varying target characteristics,most of the information fusion methods are static,the changes of confidence level of sensors and the timeliness of many characteristic indexes are not considered,but the result of the fusion recognition may be suffer from them.So a information fusion method based on fuzzy theroy and interval muti-attribute decision-making is propsed.In the method,optimization degree of information quality is defined according to the uniformity of apiece sensor to the mutual supportability of fuzzy proposition,and the characteristic recognition confidence is defined according to the interval number and multiattribute decision-making theory.Then the fusion weight vector of sensor is defined according to the two indexes,and the intelligence optimization recognition model is constructed by the method.Finally,by applying method to the target recognition,the result shows that it is effective and accurate.

关 键 词：信息融合 模糊传感器 一致性 区间型多属性决策 目标识别 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]