基于多传感器信息融合的机器人障碍检测和环境建模  

MOBILE ROBOT OBSTACLE DETECTION AND ENVIRONMENT MODELING WITH SENSOR FUSION

作　　者：邬永革[1] 黄炯[1] 杨静宇 

机构地区：[1]南京理工大学计算机系

出　　处：《自动化学报》1997年第5期641-648,共8页Acta Automatica Sinica

摘　　要：通过融合距离图象和彩色图象为移动式机器人构造二维环境模型及进行障碍检测．其中，环境模型采用了多分辨的二维网格表示结构，较好地解决了计算速度和模型精度之间的矛盾；融合算法的设计是以推广的Ｄｅｍｐａｓｅｒ－Ｓｈａｆｅｒ证据理论（ＤＳＴＥ）为基础，较好地解决了融合信息的相关性问题；所论及的系统和算法已在真实环境下经过测试，其有效性已被实验所证明．In this paper, modeling 2 D environment and road detection for mobile robot by fusing color and range image information are discussed. The environment model is constructed using multiresolution 2 D grid representation, which is proved to be a better solution to the tradeoff between accuracy and computation speed. The fusion algorithm is designed with a generalized Dempster Shafer theory of evidence ( DSTE ), which is efficient in dealing with dependent information. The multisensor system and fusion algorithm are tested in real environment and their effectiveness is proved.

关 键 词：机器人 环境建模 障碍检测 多传感器融合 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]