摘要
针对多传感器实时测量值的可靠性难以评价从而影响数据融合精度的问题,提出一种基于模糊证据理论的多传感器数据融合算法。首先将所有测量值构成的集合视为辨识框架,应用模糊数学原理引入隶属度函数,设计了多传感器之间互支持度计算的新方法,并借鉴信任分配思想将测量值转换为相应的证据,最终通过基于冲突分配的证据组合规则得到融合结果。仿真结果表明,该算法简单有效,无需传感器先验信息条件下仍可获得较高的融合精度。
Focusing on the evaluate reliability of real-time measurement of multi-sensor which directly affect the precision of data fusion,a data fusion algorithm was proposed based on fuzzy evidence theory. Firstly,the frame of discernment was obtained from the collection of measured values. By using the membership fuction based on the fuzzy mathematics principle,a new method was provided to compute the mutual support among multi-sensors. Then the concept of basic probility assignment was applied to generate the evidence. Finally the fusion result was obtained by the implementation of combination rule based on conflict assignment. The simulation result shows the simplicity and effectiveness for data fusion,higher fusion accuracy can be obtained without a priori knowledge of the sensor.
出处
《仪表技术与传感器》
CSCD
北大核心
2017年第10期118-122,共5页
Instrument Technique and Sensor
基金
国家自然科学基金项目(51475097)
工信部智能制造示范项目(工信部联装[2016]213号码)
贵州省重大基础研究(黔科合JZ字[2014]2001)
关键词
多传感器
模糊证据理论
组合规则
隶属度
数据融合
multi-sensor
fuzzy evidence theory
combination rule
membership function
data fusion
