CASIA OpenIR  > 毕业生  > 硕士学位论文
基于视觉和惯性传感器的移动机器人自定位研究
Alternative TitleResearch on Self-Localization of Mobile Robot Based on Visual and Inertial Sensors
张煌辉
Subtype工学硕士
Thesis Advisor徐德
2012-05-25
Degree Grantor中国科学院研究生院
Place of Conferral中国科学院自动化研究所
Degree Discipline控制理论与控制工程
Keyword角点提取与匹配 机器人视觉 信息融合 自定位 移动机器人 Corner Detection And Matching Robot Vision Information Fusion Self-localization Mobile Robot
Abstract本文在北京市自然科学基金项目“基于惯性和视觉传感器的移动机器人位姿估计(4082032)”的支持下,以中国科学院自动化研究所的智能轮式移动机器人AIM为实施平台,进行移动机器人基于视觉和惯性传感器的自定位研究。 本文设计了融合视觉、惯性传感器和里程计的信息进行移动机器人自身位姿估计的整体方案。移动机器人通过配备的双目视觉系统检测出工作环境中的显著角点信息,对角点进行匹配和跟踪,利用跟踪到的角点估算摄像机的位姿变换矩阵,进而推算出移动机器人的位姿变换,实现基于视觉的自定位。同时惯性传感器和里程计也能够记录移动机器人运动的姿态和位移。当获取了视觉、惯性和里程计三者的定位信息之后,利用蒙特卡洛定位方法实现移动机器人自身更加精确的定位。 提出了一种基于区域特征的Harris显著角点快速提取方法。通过候选角点的筛选过程使得角点提取的速度有了很大的提高。将精度较高的Forstner算子和Harris算子结合使用,改善了角点定位的精度。基于实验室内场景,提出了一种基于圆周区域特征的角点快速匹配方法。以基本的归一化互相关方法为基础,增加各种约束条件来保证角点匹配的正确性。设计了一种结合惯性传感器和里程计的信息以及图像之间关系的角点跟踪策略,准确的实现了角点跟踪。利用LM优化方法实现了摄像机位姿变换矩阵的求解,以及移动机器人和摄像机之间关系的推算。 分别设计了惯性传感器和里程计数据采集和处理模块,利用卡尔曼滤波器对惯性传感器数据进行处理,并对惯性传感器和里程计定位结果进行了分析。 设计了基于视觉、惯性和里程计信息的蒙特卡洛定位方法,并在实验室环境下进行了移动机器人多传感器信息融合定位实验。实验结果验证了融合定位方法的有效性。
Other AbstractThis dissertation is supported by the Beijing Natural Science Foundation under Grant 4082032, “The Pose Estimation of Mobile Robot based on Inertial and Visual Sensors”. The dissertation focuses on the research of self-localization of mobile robot based on visual and inertial sensors, using the intelligent wheeled mobile robot AIM, which is developed by the Institute of Automation, Chinese Academy of Sciences. In this dissertation, the overall positioning scheme is designed that the vision, inertial and odometer information are fused to estimate the pose of mobile robot. The mobile robot detects apparent corners in the working environment through the binocular visual system, and then matches and tracks the detected corners. The tracked corners are used to estimate the pose transformation matrix of the cameras, and then the pose transformation matrix of the mobile robot can be calculated to achieve the self-localization based on vision. At the same time, the inertial sensor and the odometer acquire the rotation and the displacement of the mobile robot. While obtaining the locating information of the vision, inertial and odometer, the Monte-Carlo localization method is utilized to locate the robot more precisely. A fast detection algorithm of Harris apparent corners based on the local features is proposed. The screening process of candidate corners is applied to speed up the corner detection. To improve the accuracy of corner position, Harris operator is combined with Forstner operator which is more precise. A fast corner matching algorithm based on circular region features is designed, and it aims at the laboratory environment. Based on the normalized cross-correlation method, several constraint conditions are employed to ensure the correctness of corner matching. Corner tracking strategy is designed to realize the corner tracking process accurately, which combines the information of inertial and odometer as well as the relationship between images. The LM optimization method is applied to calculate the pose transformation matrix of the cameras and the relationship matrix between the robot and the camera. The data acquisition and processing modules of inertial sensor and odometer are designed. The inertial data are filtered by the Kalman filter. The localization results of inertial sensor and odometer are analyzed. In the multi-sensor information fusion section, the Mento-Carlo localization based on vision, inertial and odometer is designed. The locali...
shelfnumXWLW1773
Other Identifier200928014628024
Language中文
Document Type学位论文
Identifierhttp://ir.ia.ac.cn/handle/173211/7640
Collection毕业生_硕士学位论文
Recommended Citation
GB/T 7714
张煌辉. 基于视觉和惯性传感器的移动机器人自定位研究[D]. 中国科学院自动化研究所. 中国科学院研究生院,2012.
Files in This Item:
File Name/Size DocType Version Access License
CASIA_20092801462802(2154KB) 暂不开放CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[张煌辉]'s Articles
Baidu academic
Similar articles in Baidu academic
[张煌辉]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[张煌辉]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.