北京航空航天大学学报 ›› 2018, Vol. 44 ›› Issue (5): 991-1000.doi: 10.13700/j.bh.1001-5965.2017.0325

• 论文 • 上一篇    下一篇

基于混合地图的护理机器人室内导航方法

张立志, 陈殿生, 刘维惠   

  1. 北京航空航天大学 机械工程及自动化学院, 北京 100083
  • 收稿日期:2017-05-17 出版日期:2018-05-20 发布日期:2018-05-29
  • 通讯作者: 陈殿生 E-mail:chends@163.com
  • 作者简介:张立志,男,博士研究生。主要研究方向:机器人移动导航;陈殿生,男,博士,教授,博士生导师。主要研究方向:助老助残机器人、仿生机器人、机电智能控制技术;刘维惠,女,博士研究生。主要研究方向:机械臂运动控制。
  • 基金资助:
    北京市科技计划重大项目课题(D141100003614002)

Care robot indoor navigation method based on hybrid map

ZHANG Lizhi, CHEN Diansheng, LIU Weihui   

  1. School of Mechanical Engineering and Automation, Beijing University of Aeronautics and Astronautics, Beijing 100083, China
  • Received:2017-05-17 Online:2018-05-20 Published:2018-05-29

摘要: 护理机器人在室内三维结构化环境下进行导航时,面临着三维建图计算量大且地图中缺乏语义信息的缺点。提出了基于点和平面特征的混合地图构建方法,结合点和平面在地图构建中的优势,并基于该混合地图搭建室内导航系统。首先,快速地提取特征点和特征平面,使用解释树的方法进行数据关联,并使用平滑建图工具构建因子图,进行机器人位姿和路标的联合优化,改进并更新混合地图。然后,搭建室内导航系统,实现了三维障碍物检测、路径规划与运动控制。最后,在走廊环境下进行了室内导航实验,并以由激光雷达构建的二维栅格地图为参考,分析了地图构建效果和机器人定位精度,证明了基于混合地图的室内导航系统在室内结构化环境下的优势。

关键词: 护理机器人, 室内导航, 同步定位与建图, 数据关联, 障碍物检测, 路径规划

Abstract: When the care robot is navigating in the indoor 3D structured environment, it is faced with the disadvantage of the large computational cost for map building and the lack of semantic information in the map. This paper presents a hybrid map building method based on point and plane features, which combines the advantages of point and plane features in the map building. Furthermore, an indoor navigation system is built based on the proposed hybrid map. First, point and plane features are fast extracted, and then data association is achieved using the interpretation tree approach. The smoothing and mapping tool is utilized to construct the factor graph and jointly optimize robot poses and landmarks, and the hybrid map is refined and updated. Second, the indoor navigation system is built, which implements the 3D obstacle detection, path planning and motion control. Finally, the indoor navigation experiments were carried out in a corridor environment. With the 2D occupancy grid map constructed by laser as the reference, the performance of map building and robot localization accuracy were analyzed, which proves that the indoor navigation system based on hybrid map shows its advantages in indoor structured environments.

Key words: care robot, indoor navigation, simultaneous localization and mapping, data association, obstacle detection, path planning

中图分类号: 


版权所有 © 《北京航空航天大学学报》编辑部
通讯地址:北京市海淀区学院路37号 北京航空航天大学学报编辑部 邮编:100191 E-mail:jbuaa@buaa.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发