摘要
基于最优估计理论,提出了一种多层传感器的复合融合估计算法.该方法采用了由真实传感器和虚拟传感器组成的复合融合结构,以解决测量噪声干扰下参数估计问题.在该结构中,第1层物理真实传感器测量所得数据,按照加权融合算法,经过多层虚拟传感器的递推融合得到多层融合后的估计值.虚拟传感器的层数增加时,融合后总的估计误差将减少.虚拟传感器是由算法和软件实现的,不会增加测量系统物理结构的复杂性.计算机仿真结果表明:此方法在减少测量误差方面比传统的直接估计方法有明显的改善.
Based on the theory of optimal estimation, this paper proposed a compound fusion estimation algorithm of multi-layer sensors. The compound fusion structure, made up of real sensors and visual sensors, was used to solve the parameter estimation problems in noise environment. In this structure, the data from the first layer real physical sensors are processed recursively by multi-layer virtual sensors under a weighing fusion algorithm, so as to obtain their estimation value. The total number of estimation errors from data fusion will be decreased when layers of virtual sensors are increased. Since the virtual sensors are realized by software and algorithm, this true virtual data fusion method with multi-layers will not increase the complexity of the physical structure of measurement systems. Simulation results showed that the method is more efficient in reducing measurement errors than the traditional centralized algorithm.
出处
《湖南大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2004年第5期59-61,共3页
Journal of Hunan University:Natural Sciences
基金
国家自然科学基金资助项目(60272051)
关键词
最优估计
集员估计
多层信息融合
optimal estimation
set-member estimation
fusion of multi-layer information