中国科学院沈阳自动化研究所机构知识库
Advanced  
SIA OpenIR  > 机器人学研究室  > 学位论文
题名: 复杂环境下同步定位与地图创建研究
其他题名: Study on Simultaneous Localization and Mapping in Complex Environments
作者: 孙荣川
导师: 马书根 ; 王越超
分类号: TP242
关键词: 移动机器人 ; 导航 ; 同步定位与地图创建 ; 离散状态估计 ; 数据分割
索取号: TP242/S97/2010
学位专业: 机械电子工程
学位类别: 博士
答辩日期: 2010-11-16
授予单位: 中国科学院沈阳自动化研究所
学位授予地点: 中国科学院沈阳自动化研究所
作者部门: 机器人学研究室
中文摘要: 近年来,随着机器人技术的飞速发展,机器人正代替人类走向越来越多的工作场合。在机器人具备的众多能力里面,自主性是其中极其重要的一个。具有自主能力的移动机器人可以在无人指导的情况下,在未知的环境中独立完成事先给定的任务。机器人自主性的一个前提就是必须知道“我在哪”和“周围环境是什么”,也就是定位与地图创建能力。机器人只有在具备定位和地图创建能力的情况下才能实现其它更高层的智能性算法。    本文的研究工作主要源于国家“863”高技术计划重点项目子课题“废墟洞穴搜救机器人研制”。以灾难救援为背景,研究机器人在大尺度复杂未知环境、没有全局位置信号(如GPS等)的情况下,利用其搭载的传感器给自身定位并创建周围环境地图的问题,为将来研制废墟内部自主式搜救机器人做理论预研。论文围绕机器人在复杂未知环境中的定位与地图创建问题,系统地研究了SLAM领域里的不确定性信息处理、复杂地图创建、数据关联和大尺度环境下的实时性技术。具体研究内容如下:    1) EKF-SLAM的一致性问题。传统的基于特征地图的SLAM算法在长时间运行之后,由机器人位姿与地图组成的状态并不收敛。即使在最简单的仿真环境下,状态估计也会出现不一致性现象。关于这个问题,目前学术界还没有有效的解决办法。本文分析了这种现象产生的原因,从卡尔曼滤波器所优化的目标函数出发,提出利用相同线性展开点,对相同物理量在不同时刻的观测模型采用相同的线性拟合形式,改进了EKF-SLAM的更新过程,从而提高了它的一致性性能。该方法可以用于基于特征地图的EKF-SLAM、混合式地图创建等一类算法中用于融合不确定性信息,提高它们的方差一致性。    2) 同步定位与采样环境地图创建。针对复杂未知环境下移动机器人的定位与地图创建问题,本文提出了同步定位与采样环境地图创建(SLASEM)算法。相对于已有的复杂环境下的SLAM算法间接地创建复杂环境地图、并且只能得到状态的次优估计,SLASEM能够在滤波器的一个更新步骤中给机器人定位并创建描述复杂环境的地图,这使SLASEM能够得到概率意义上的最优估计,其所得到的机器人位姿估计与地图更精确。这部分的研究分为采样环境地图的更新与维护、复杂环境中的数据关联以及大尺度环境下的实时性技术:    a) 采样环境地图的更新与维护。针对采样环境地图更新、冗余粒子约简以及地图拓扑结构奇异等一系列关键性的科学问题,分别提出了基于代数距离和符号正交距离的观测模型、基于几何约束的冗余粒子约简算法,以及拐角约束算法,有效地创建并维护了地图。    b) 复杂环境中的数据关联。针对采样环境地图与实际环境之间不具备一一对应性、已有的数据关联算法都不再适用的问题,提出利用点集合之间的广义距离函数来衡量测量值和预测地图中物体之间的相似性,基于此距离函数提出了能适用于SLASEM的数据关联算法。针对多重数据关联问题,提出利用复杂环境结构信息的改进数据关联算法。    c) 大尺度环境下的实时性技术。针对大尺度复杂环境下的SLAM问题,将分治法引入到SLASEM中,提出了基于分治法的SLASEM算法。该方法能使机器人在大尺度的环境中实时地定位,并创建复杂地图。针对其中的局部地图融合这一关键性的问题,提出了角点约束更新、基于符号正交距离的更新和粒子约简算法。
英文摘要: With the rapid development of robotics in recent years, robots have been put into more and more workspace to replace human beings. The autonomous ability is one of the most important abilities of robots. An autonomous robot can accomplish tasks in unknown environments individually without any guidance. For a robot, autonomy means that it must know “Where am I” and “What is the map”, which are the problems of localization and mapping. The ability of localization and mapping is indispensable for an autonomous robot to execute other higher level intelligent algorithms. This paper is mainly supported by the project of Research of the Rescue Robot in Ruins and Caves, which is a sub-project of National High Technology Research and Development Program 863. The research focuses on the problem how to localize the robot and map the environment when it moves in a large scale unknown complex environment without global position information (e.g. GPS). In future, the results of this research will be used in the rescue robot in ruins and caves. To solve the problem of localization and mapping in a complex unknown environment, we study on the SLAM problem systematically from the aspects of uncertainty processing, complex map building, data association, and the technologies of realizing the SLAM algorithm in real-time. The contents of this thesis can be summarized mainly as follows. 1) The consistency problem of EKF-SLAM. The traditional feature-based EKF-SLAM algorithms have a problem that the state which is composed of the robot’s pose and the map does not converge after running for a long time. Even in a simple simulated environment, the estimated state will still have inconsistency. The inconsistency problem has not yet been solved in SLAM realm. This paper analyzes the factors that cause the inconsistency, and proposes an improved algorithm from the point of the objective function that a Kalman filter minimizes. This algorithm linearizes observation models at different time instants corresponding to a same physical quantity at the same point, so that all the linear approximations of the observations have the same formula. In this way, the consistency of EKF-SLAM algorithms is improved. The proposed algorithm can be used in feature‐based EKF-SLAM and hierarchical mapping algorithms to improve their covariance consistency.  2) Simultaneous localization and sampled environment mapping. To localize the robot and map a complex environment, this paper proposes the algorithm of simultaneous localization and sampled environment mapping (SLASEM). This algorithm can estimate the robot’s pose and build a complex map in one filtering stage, which enables SLASEM an optimal result from the point of probability. Compared with the other algorithms which build the complex map indirectly and achieve an sub‐optimal result, SLASEM can estimate the robot’s pose and build a complex map more accurately. In this algorithm, a series of scientific problems are solved, including the update and management of the SEM, data association in complex environments, and real‐time technologies in large scale environments. a) Update and management of the SEM. In this field, three key scientific problems are solved. i) A new observation model using algebraic distance or signed orthogonal distance is proposed to update the map. ii) A method based on geometric constraint is proposed to reduce the redundant environment samples. iii) The corner constraint algorithm is proposed to improve the map’s topological consistency.  b) Data association in complex environments. Because there are no one-to-one correspondences between the environment samples in the map and the ones in real environments, traditional data association algorithms cannot be used in SLASEM. This paper proposes to use a general distance function between two point sets to value their similarity. Based on this function, a data association algorithm is proposed. An improved algorithm which utilizes the information of the environment’s topological structure is also proposed to solve the multi-association problem. c) Real-time technologies in large scale environments. The divide and conquer SLASEM algorithm is proposed to meet the requirements of the problem that a mobile robot navigates autonomously in a large scale unknown environment in real time. This algorithm has a computation complexity of O(N), which enables it to be performed on-line in large scale unknown environments. Sub-map joining, which is a key problem in this research, is solved by using corner point update, signed orthogonal distance update, and samples reduction sequentially.
语种: 中文
产权排序: 1
内容类型: 学位论文
URI标识: http://ir.sia.cn/handle/173321/9378
Appears in Collections:机器人学研究室_学位论文

Files in This Item:
File Name/ File Size Content Type Version Access License
复杂环境下同步定位与地图创建研究.pdf(2826KB)----限制开放 联系获取全文

Recommended Citation:
孙荣川.复杂环境下同步定位与地图创建研究.[博士 学位论文 ].中国科学院沈阳自动化研究所 .2010
Service
Recommend this item
Sava as my favorate item
Show this item's statistics
Export Endnote File
Google Scholar
Similar articles in Google Scholar
[孙荣川]'s Articles
CSDL cross search
Similar articles in CSDL Cross Search
[孙荣川]‘s Articles
Related Copyright Policies
Null
Social Bookmarking
Add to CiteULike Add to Connotea Add to Del.icio.us Add to Digg Add to Reddit
所有评论 (0)
暂无评论
 
评注功能仅针对注册用户开放,请您登录
您对该条目有什么异议,请填写以下表单,管理员会尽快联系您。
内 容:
Email:  *
单位:
验证码:   刷新
您在IR的使用过程中有什么好的想法或者建议可以反馈给我们。
标 题:
 *
内 容:
Email:  *
验证码:   刷新

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

 

 

Valid XHTML 1.0!
Copyright © 2007-2016  中国科学院沈阳自动化研究所 - Feedback
Powered by CSpace