Towards autonomous 3D exploration using surface frontiers
Towards autonomous 3D exploration using surface frontiers
复制标题
使用表面前沿进行自主 3D 探索
DOI：
--
复制
发表时间：
2016
期刊：
影响因子：
--
通讯作者：
Danwei W. Wang
中科院分区：
文献类型：
--
作者：
P. Senarathne;Danwei W. Wang
探索环境以生成3D地图是任何预期用于搜索和救援或安全相关任务的自主机器人的核心组件。考虑到三维网格地图中的表面体素信息对于许多任务来说是足够的，在执行映射使命之后，例如搜索或救援的运动规划，本文提出了一种由表面边界而不是自由空间边界引导的探索策略，以主动映射环境中的对象表面。3D网格地图中的表面边界是邻近未映射空间的表面边界体素。详细讨论了从3D占用网格检测表面边界，并且提供了使用检测到的表面边界来对视图配置进行采样以进行探索的策略。在模拟和真实的室内环境中进行的探索实验验证了表面边界是自主3D探索的自由空间边界的有效和高效的替代方案。
Exploring environments to generate 3D maps is a core component of any autonomous robot expected to operate in search and rescue or security related missions. Considering that information about surface voxels in a 3D grid map is sufficient for many tasks, following a mapping mission, such as motion planning for searching or rescue, this paper presents an exploration strategy guided by surface frontiers rather than free-space frontiers to actively map the object surfaces in the environment. Surface frontiers in a 3D grid map are surface boundary voxels neighboring unmapped space. Detecting surface frontiers from a 3D occupancy grid is discussed in detail and a strategy to sample view configurations using the detected surface frontiers for exploration is provided. Exploration experiments conducted in both simulated and real indoor environments validate that surface frontiers are a valid and an efficient alternative for freespace frontiers for autonomous 3D exploration.
影响因子：
3.5
作者：
Hornung, Armin;Wurm, Kai M.;Burgard, Wolfram
通讯作者：
Burgard, Wolfram