基于随机自适应方法的多传感器融合算法  

MULTI-SENSOR FUSION ALGORITHM BASED ON RANDOM ADAPTIVE METHOD

作　　者：王茜 王颖超[1] 曹菲[2] Wang Qian;Wang Yingchao;Cao Fei(Public Department of Hebei Academy of Fine Arts,Shijiazhuang 050700,Hebei,China;Xidian University,Xi’an 710077,Shaanxi,China)

机构地区：[1]河北美术学院公共课部,河北石家庄050700 [2]西安电子科技大学,陕西西安710077

出　　处：《计算机应用与软件》2020年第4期233-239,共7页Computer Applications and Software

基　　金：国家自然科学基金项目(61571460);河北省众创空间相关问题研究项目(JRSHZ-2016-01091)。

摘　　要：针对基于熵理论的贝叶斯信息融合技术需要进行无穷区间的积分运算,容易出现数值不稳定的问题,提出一种基于随机自适应方法的多传感器融合算法。利用传感器测量值之间的差值自适应地建立传感器的后验概率分布模型;结合互信息的理论实时识别和剔除伪测量值,避免求熵时的积分计算;将该方法分别应用于集中式融合方案和分布式融合方案中得到了两种新的数据融合方法。仿真实验结果表明,在存在伪测量值的情况下,该算法性能明显优于一般的贝叶斯融合方法。Aiming at the problem that Bayesian information fusion technology based on entropy theory needs integral operation of infinite interval and easy numerical instability,a multi-sensor fusion algorithm based on random adaptive method is proposed.It established the posterior probability distribution model of the sensor by using the difference between the measured values of the sensor.Then,the theory of mutual information was combined to identify and eliminate the pseudo-measured values in real time,so as to avoid the integral calculation in the entropy calculation.Finally,two new data fusion methods were obtained by applying the method to the centralized fusion scheme and the distributed fusion scheme respectively.The simulation results show that the performance of the proposed algorithm is better than the general Bayesian fusion method in the presence of pseudo-measurement values.

关 键 词：多传感器信息融合 贝叶斯方法 伪数据 自适应建模 互信息 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]