摘要
为解决火灾探测问题,采用基于推理的信息融合方法,分析了Dempster-shafer证据理论用于多传感器数据融合的基本概念和理论,并将它用于火灾探测的数据融合中。在数据处理方法上,结合具体应用背景,借助专家数据库知识,构造初始概率分配函数,产生每条证据的信度函数分配。在融合规则上,采用改进的融合规则对多条证据进行数据融合。融合结果鲁棒性强,有效地解决了传感器数据失效及证据冲突等问题。
A method of data fusion based on reasoning is applied to solve the problem of fire detection.This paper analyzes the basic concepts and theories of Dempster-shafer evidence theory which is applied to multi-sensor data fusion in the field of fire detection.In the way of data processing,a method of constructing the initial probability distribution function is developed by combining the specific application background with the knowledge of expert database in order to produce the reliability distribution of each evidence.In terms of the fusion rules,the improved integration rule is used to fuse a number of evidences.The fusion results are strongly robust and effectively solve the problem of unreasonable fuses caused by invalidation of sensor data and the conflicts between evidences.
出处
《重庆理工大学学报（自然科学）》
CAS
2011年第1期44-48,共5页
Journal of Chongqing University of Technology：Natural Science
基金
重庆市科委科技攻关项目(CSTC
2009AC2068)
关键词
信息融合
D-S证据推理
基本概率赋值函数
融合规则
专家数据库
data fusion
Dempster-shafer evidence theory
basic probability assignment function
combination rule
expert database
