摘要
近年来,多传感器数据融合的研究因其优异的性能而受到广泛关注.为了缓解高冲突数据对融合的影响,本文提出了一种基于皮尔逊系数和信息熵的多传感器数据融合方法.首先,引入皮尔逊系数来衡量证据碎片之间的冲突程度,得到证据的信度;同时,采用信息熵来度量证据的不确定性.然后,根据得到的信息熵对证据的可信度进行修正,进一步得到加权平均证据.最后,根据Dempster的组合规则对加权证据进行数据融合.数值算例结果表明,该方法具有较好的鲁棒性,更适合于实际场景.在目标检测和故障诊断中的应用结果表明,该方法的准确率分别高达98.09%和91.94%,证明了该方法的有效性和可行性.
In recent years,the research of multi-sensor data fusion has received extensive attention due to its excellent performance.To alleviate the impact of high-conflict data on fusion,a multi-sensor data fusion method based on Pearson coefficient and information entropy is proposed in this paper.Firstly,the Pearson coefficient is introduced to measure the degree of conflict between two evidence shreds,and the reliability of evidence is obtained.Meanwhile,information entropy is adopted to measure the uncertainty of evidence.Then,based on the obtained information entropy,the credibility of the evidence is corrected,and the weighted average evidence is then further available.Finally,data fusion is performed according to Dempster′s combination rules on the weighted evidence.Numerical examples results show that the proposed method has better robustness and is more suitable for real scenarios.In addition,the results of the applications on target detection and fault diagnosis show that the accuracies are as high as 98.09%and 91.94%respectively,which demonstrates the effectiveness and feasibility of the proposed method.
作者
陶洋
祝小钧
杨柳
TAO Yang;ZHU Xiao-jun;YANG Liu(College of Communication and Information Engineering,Chongqing University of Posts and Telecommunications,Chongqing 400065,China)
出处
《小型微型计算机系统》
CSCD
北大核心
2023年第5期1075-1080,共6页
Journal of Chinese Computer Systems
基金
国家重点研发计划项目(2019YFB2102001)资助
国家自然科学基金项目(61801072)资助
重庆市自然科学基金项目(CSTC2018jcyjAX0344)资助
重庆市教委项目(KJQN202000641)资助。
关键词
DS证据理论
皮尔逊相关系数
信息熵
数据融合
DS evidence theory
pearson coefficient
information entropy
multi-sensor data fusion
