项目名称: 模仿人的环境认知机理的机器人室内地图自主构建研究
项目编号: No.61203330
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: 吴皓
作者单位: 山东大学
项目金额: 24万元
中文摘要: 根据人的区域功能分割、对象语义描述的环境认知模式,针对机器人日趋扩大的活动范围和种类繁多的服务对象特征,结合基于RFID和QR code的人工物/路标,模仿人的对象管理、对象移动、对象操作的三级服务任务完成模式,研究机器人完全自主地构建三级室内地图的策略和方法。该研究使服务机器人对环境信息的掌握不仅停留在面向导航的几何结构信息方面,还增加了局域空间的特征描述信息和物品操作信息,及楼层、走廊、房间和物品功能及归属关系的语义信息,以满足层次化智能导航的要求。模仿人的基于人工路标导航的大范围空间认知机理和基于物品属性和关联关系的小范围空间功能描述机理,构建基于免疫网络算法的大范围认知语义拓扑图和基于谱聚类算法的小范围功能语义拓扑图。同时基于描述逻辑推理实现感知、定位、语义、上下文、对象管理各部分功能的深度整合,支持与服务相关的"语义层"和"区域层"路径规划,为机器人智能化的服务任务实现奠定基础。
中文关键词: 语义;人工标签;三级地图;地图更新;智能服务
英文摘要: According to the human's environment cognitive mode with region division function and object semantic description, and aiming to the service robot's larger moving range and kinds of service objects, the strategy and method of the three-level indoor map building using the artificial signpost/object mark based on RFID and QR code is proposed to imitate the human's three-level service task completing mode with object management, moving and operations. The three-level maps make the robot not only achieving the plane structure of the environment for navigation, but also enriching the information of typical properties and objects operation of local space, and semantic information of describing the function, relationship & ascription of the floors, corridors, rooms and objects, which are used to meet the requirements of the hierarchical and intelligent navigation. Imitating the human's space cognitive mechanism by navigation of artificial signpost in large-scale environments and the space functional description mechanism by the object property & relationship in small-scale environments, large-scale cognitive semantic topology map based on immune network algorithm and small-scale functional semantic topology map are constructed. Besides, the functions of perceiving, localization, semantic map, context and object manage
英文关键词: semantic;artificail label;three hierarchical map;map update;intelligence service