基于红外路标的移动机器人室内定位研究
文献类型:学位论文
作者 | 胡鹏 |
学位类别 | 工学硕士 |
答辩日期 | 2010-05-31 |
授予单位 | 中国科学院研究生院 |
授予地点 | 中国科学院自动化研究所 |
导师 | 李成荣 |
关键词 | 移动机器人 室内定位 红外路标 单应矩阵 Mobile Robot Indoor Localization Infrared Landmark Homography |
其他题名 | Indoor Localization of Mobile Robot Based on Infrared Landmarks |
学位专业 | 模式识别与智能系统 |
中文摘要 | 随着人们生活需求的提高和科技水平的进步,室内移动机器人的应用将越来越广泛。在自主移动机器人的研究领域中,机器人自定位是一个非常重要的研究方向,是实现机器人导航和其它智能行为的前提。本文针对运行在室内环境下的移动机器人的全局定位问题开展研究,提出了一种基于红外路标的扩展单应矩阵定位方法,主要内容和贡献如下: 本文首先采用基于平面模板的摄像机标定方法实现广角摄像机的参数标定,并利用畸变参数对图像进行畸变校正。 针对大范围室内环境的移动机器人的自定位问题,本文设计了一种新的红外点阵路标,该红外点阵路标为红外LED 制作的3*3型的主动式红外路标,红外点阵外围特定的四点用于路标子坐标的建立以及摄像机定位计算,红外点阵中其它的点则采用二进制编码的方式计算路标的ID值。该路标具有易检测、易识别、利于定位计算、制作方便等特点。 本文提出了移动机器人室内定位的平面假设:路标平面与移动平台运动平面平行,固定在移动平台上的单目摄像机跟随移动平台运动,且只有平移和旋转两种运动模式。然后通过分析传统的单应矩阵定位和基于共面PnP问题的定位方法,本文提出了一种新的扩展单应矩阵定位方法。该方法利用安装在移动机器人上的单目摄像机实时检测贴附于天花板上的主动红外路标,通过假设单应矩阵不变来计算出虚拟路标的位置,由虚拟路标与实际路标之间的位姿差异推算出摄像机的位姿变化,从而达到移动机器人定位的目的。该方法计算简单、处理快速、定位精度高、无累计误差,与本文设计的红外路标结合,构成一套完整的、具有实用性的移动机器人室内定位系统。 本文最后通过定位系统的具体实现以及实验设计与结果分析验证了该方法的准确性和快速性等特点,给出了需进一步开展的工作,以便将此系统转化为实用的产品。 |
英文摘要 | With the improvement of technology, indoor mobile robot will be widely applied to our daily life. As the prerequisite to realize navigation and other intelligent actions, robot localization is one of the most important research branches of autonomous mobile robot. Considering global localization of indoor mobile robot, an extended homography localization method based on infrared landmarks is proposed. Contributions of this paper are summarized as follows: The method of camera calibration based on planar pattern is used to calibrate the wide-angle camera. Then, the image is corrected by distortion parameters. In order to resolve the problem of precise robot localization in a large indoor area, this paper proposes a new kind of active infrared landmarks, which are a series of dot-matrix landmark (3*3) made by infrared LED. In each landmark, certain four external points are used to establish the sub-coordinate of the landmark and calculate the pose of the camera, the remaining points are used to calculate the ID value of the landmark by binary coding. The landmark has many advantages, such as easy-detection, easy-recognition, easy-localization and so on. A majority of indoor mobile robot can be reasonably considered to perform under such a plane assumption: the plane of the landmarks is parallel to the moving plane of the robot; the camera, fixed on the robot, has two movement patterns – translation and rotation. After analyzing the traditional localization method based on homography and PnP problem, this paper introduces a new extended homography localization method. A camera fixed on the robot is used to capture the image of infrared landmarks attached to the ceiling. The location of virtual landmark is calculated by keeping the homography to be constant, then, the position and orientation change of the camera is calculated through comparing the location of virtual landmark and real landmark. Besides the advantage of high effectiveness, accumulated error is avoided in the algorithm. Combined with the infrared landmarks, a complete and practical localization system for indoor mobile robot can be formed. This paper put the localization system into experiment, and the results show that the method is fast and accurate as expected. Finally, further work of this paper is introduced for future development of practical localization system. |
语种 | 中文 |
公开日期 | 2015-09-08 |
其他标识符 | 200728014629032 |
源URL | [http://ir.ia.ac.cn/handle/173211/7523] ![]() |
专题 | 毕业生_硕士学位论文 |
推荐引用方式 GB/T 7714 | 胡鹏. 基于红外路标的移动机器人室内定位研究[D]. 中国科学院自动化研究所. 中国科学院研究生院. 2010. |
入库方式: OAI收割
来源:自动化研究所
浏览0
下载0
收藏0
其他版本
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。