项目名称: 基于点集配准的栅格地图拼接方法研究
项目编号: No.61203326
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 自动化学科
项目作者: 祝继华
作者单位: 西安交通大学
项目金额: 24万元
中文摘要: 作为多移动机器人系统协同创建地图中的一项关键技术,地图拼接是移动机器人领域一项极富挑战性的课题。本项目旨在充分应用点集配准理论,对栅格地图拼接问题展开深入地研究,进而设计出一套具有高精准度和强鲁棒性的地图拼接算法。以申请人已有的地图创建研究工作为基础,结合现有的点集配准研究成果,通过分析将不同类型的栅格地图拼接问题转化为相应的部分匹配点集配准问题,并讨论如何设计对应的局部收敛的点集配准算法;在此基础之上,通过分析点集配准问题的全局收敛条件,进一步讨论如何设计两类全局收敛的算法:基于配准初值分析的点集配准算法和基于粒子滤波的点集配准算法,以获得精确且可靠的栅格地图拼接结果。本项目所研究的算法框架具有较强的通用性,拟设计的方法可解决不同类型的栅格地图拼接问题。该项目的实施可为大规模栅格地图的创建工作提供可靠的技术手段,相应的研究成果可应用于太空、深海以及矿井等极端环境下的探测与开发。
中文关键词: 同步定位与地图创建;多移动机器人系统;栅格地图拼接;点集配准;
英文摘要: As one of the key techniques in cooperative mapping by the multiple mobile robot system, map merging is a challenging issue in the field of mobile robot. Therefore, this project will make deep research on grid map merging problem and attempt to propose the accurate and robust solutions with the theory of point set registration. Based on our research work on mapping and the existing achievements of point set registration, we will analyze different map merging problems and transform them into appropriate partial registration problems. Then, we will discuss how to design the corresponding local convergent registration algorithms for each of the partial registration problems. On the basis above, we will analyze the global convergent condition of registration problem and discuss how to design two kinds of global convergent algorithms,namely the point set registration algorithm based on the estimation of initial parameters and the point set registration algorithm based on particle filter. Both of them are expected to get the reliable and accurate results for different map merging problems. The proposed approach of this project can solve different grid map merging problems, and thereby it will be a general framework for grid map merging. Besides, this project will play an important role in the application of grid mappi
英文关键词: Simultaneous localization and mapping;Multiple robot system;Grid map merging;Point set registration;