基于修正狄克逊准则的多传感器融合算法  

A Multi-Sensor Fusion Algorithm based on Modified Dixon Criterion

作　　者：陈亚斌[1] 王亚刚[1] 周代仝 

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093 [2]上海煊荣自动化科技有限公司,上海201106

出　　处：《通信技术》2014年第10期1178-1182,共5页Communications Technology

基　　金：上海市一流学科(系统科学)项目资助(No.XTKX2012)~~

摘　　要：采用多传感器测量某一观测值时,对测量值进行融合处理,可以实现比较准确的估计,从而做出正确的决策。但是,测量值中存在的恶劣异常值,而直接剔除异常值会损害融合算法。针对这两点提出了基于修正狄克逊准则的多传感融合算法,该算法不但有效地去除恶劣异常值,而且保证了融合算法的可行性。它不但在理论上得到论证,而且实例仿真分析也证明其效果显著,可以有效地控制被控对象,提高了系统的控制精度和准确性。Wiht multiple sensors to measure an observed value, fusion processing should be usually done on the measured value, thus to achieve relatively precise estimation and rational decision. However, in measured value always exists some malicious abnormal value, and direct eliminaton of abnormal value would damage fusion algorithm. Aiming at these two points and based on modified Dixon criterion, a multi-sensor fusion algorithm is proposed, which could effectively remove malicious abnormal value and ensure the feasibility of fusion algorithm. This algorithm is verified both by theory and simulation. The proposed algorithm can effectively control objects and improve the precision and accuracy of system.

关 键 词：多传感器 修正狄克逊准则 融合算法 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]