基于信息融合的管道机器人定位控制研究  

Localization Technique of Pipeline Robot Based on Multi-sensor Data Fusion

作　　者：李军远[1] 陈宏钧[1] 张晓华[1] 邓宗全[1] 

机构地区：[1]哈尔滨工业大学电气工程系,哈尔滨150001

出　　处：《控制与决策》2006年第6期661-665,共5页Control and Decision

基　　金：国家863计划项目(2001/AA422410)

摘　　要：提出了低频电磁波、视觉、计程轮这3种管道机器人定位方法;分析了3种传感器定位精度、作用范围和工作速率的特点;在考虑3种定位方法精度条件下,给出了传感器的优先权系数计算方法与基于优先权系数的D-S证据融合信度函数;通过D-S证据融合规则得到融合证据,解决了多传感器定位中决策传感器的选择问题,以实现管道机器人准确、快速地定位.Low frequency electromagnetic wave method, visual recognition method and odometer method are presented to achieve the pipeline robot＇s localization. The precision and data range of these three methods are analyzed. According to different precision, the sensor priority factor is calculated, based on which the belief function is obtained. Through D-S data fusion rules using the belief function, the belief function assignment is obtained to solve the multi-sensors selection problem and to have the accuracy localization data of pipeline robot.

关 键 词：管道机器人 低频电磁波 定位 信息融合 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]