机器人超高分辨率立体网格导航地图建模研究

Robotic Navigation Map Construction with Ultra-High Resolution Volumetric Grid

针对机器人定位和导航的具体应用需求,提出一种供"机器人"使用的超高分辨率立体网格导航地图的空间建模方法。创新性地将地理空间立体网格模型与机器人导航地图相结合,分析其空间跨度、网格单元和数据组织的一致性。详细分析了机器人导航地图的特点、数据来源和分类,及其与SLAM成图的区别和联系。以超高分辨率立体网格单元作为数据组织单元,将机器人导航地图空间数据集进行离散化,提出了机器人周围空间环境数据的集成管理方法,并进行了优化。利用超高分辨率地理空间立体网格模型对机器人周围环境空间网格进行统一编码和索引,实现了机器人周围环境的空间建模,为移动机器人定位和导航提供了统一的环境地图模型。并利用点云数据进行了试验,进一步论证了机器人超高分辨率立体网格导航地图建模的可行性。

According to the requirements of specific application of robotic localization and navigation, a spatial modeling method of the robotic navigation map with ultra-high resolution volumetric grid model is proposed. The geographic spatial three-dimensional grid model is combined with the robotic navigation map and the consistency of spatial span, grid cell and data organization is analyzed. The characteristics of the robotic navigation map, the source and classification of the data, and the difference and relationship between SLAM are analyzed in detail The robotic spatial data sets are discretized with the ultra-high resolution volumetric grid units as the data organization units. The integration management method of the spatial environment data is proposed and optimized. The unified coding and indexing of the environment space grid around the robot is conducted using the ultra-high resolution volumetric grid model, which achieves the spatial modeling of the environment around the robot and provides a uniform environment map model for the localization and navigation of mobile robot. The point clouds data from the robotic vision sensor demonstrates the feasibility of robotic navigation map construction with ultra-high resolution volumetric grid.