中国机械工程 ›› 2022, Vol. 33 ›› Issue (02): 194-201.DOI: 10.3969/j.issn.1004-132X.2022.02.009

• 智能制造 • 上一篇    下一篇

大视场室内移动机器人高精度动态定位方法

田明锐1,2;杨皓1;胡永彪1   

  1. 1.长安大学道路施工技术与装备教育部重点实验室,西安,710064
    2.广西柳工机械股份有限公司,柳州,545006
  • 出版日期:2022-01-25 发布日期:2022-02-18
  • 作者简介:田明锐,男,1984年生,博士、工程师。研究方向为工程机械机器视觉理论及技术。E-mail:tianmingrui1020@163.com。
  • 基金资助:
    中国博士后科学基金(2018M643551);
    陕西省自然科学基础研究计划(2020JM-240);
    中央高校基本科研业务费专项资金(300102258305)

Large-field and High-precision Dynamic Positioning Method of Indoor Mobile Robots

TIAN Mingrui1,2  ;YANG Hao1;HU Yongbiao1   

  1. 1.Key Laboratory of Road Construction Technology and Equipment of MOE,Chang'an University,Xi'an,710064
    2.Guangxi Liugong Machinery Co.,Ltd.,Liuzhou,Guangxi,545006
  • Online:2022-01-25 Published:2022-02-18

摘要: 高精度室内定位是移动机器人实现自主导航、运动控制和协同作业等任务的前提条件。利用单目视觉定位原理,提出了一种针对大视场的室内移动机器人绝对定位方法。该方法适用于室内多移动机器人同步动态定位,解决了大视场环境下多相机非线性标定和视场融合问题,相机标定精度提高了52.6%。为提高移动机器人动态定位精度,设计了基于光信标的高速定位算法,提高了定位系统的图像处理速度,并对定位系统动态误差进行了预测和补偿。对定位方法中各算法进行实验验证,结果表明提出的方法可以对多个室内高速移动机器人进行高精度动态定位,当机器人移动速度小于3.2 m/s时,系统位置平均动态定位误差低于5 cm,姿态角平均动态定位误差低于0.6°。

关键词: 室内定位, 移动机器人, 单目视觉, 图像处理, 动态定位

Abstract: High-precision indoor positioning process was the basis and prerequisite for mobile robots to fulfill tasks such as autonomous navigation, motion control, and cooperative operation. a large-field and absolute positioning method was proposed for indoor mobile robots based on the principle of monocular vision positioning. This method was suitable for the synchronous dynamic positioning of indoor multi-mobile robots, the problems of multi-camera nonlinear calibration and field fusion in the large-field environment were solved, and the camera calibration accuracy was improved by 52.6%. A high-speed positioning algorithm was designed based on optical beacons to improve the dynamic positioning accuracy of the mobile robots, where the image processing speed of the positioning system was improved. Then the dynamic errors of the positioning system were predicted and compensated. The experimental verification of each algorithm shows that the proposed method may perform high-precision dynamic positioning of multiple indoor high-speed mobile robots. When the robot moving speed is less than 3.2 m/s, the average dynamic positioning errors of the system position are less than 5 cm, and the average dynamic positioning errors of the system attitude angle are less than 0.6°.

Key words: indoor positioning, mobile robot, monocular vision, image processing, dynamic positioning

中图分类号: