Large-scale Map Building by a Mobile Robot Using High-dimensional Visual Sensors

本文从“高维传感器”（如视觉传感器）的新视角研究了同步定位与地图构建问题。新技术的几个组成部分，包括一个通用的和紧凑的地标（视觉特征），视觉特征数据库，以及“地图匹配SLAM”已经开发，统一到一个新的框架，并通过实验在室内和室外环境中进行评估。

The problem of SLAM (simultaneous localization and mapping) has been approached from a novel perspective of "high-dimensional sensor" (e.g. visual sensor) in this study. Several building blocks of the novel technique, including a universal and compact landmark (visual feature), visual feature database, as well as "map-matching SLAM" have been developed, united into a novel framework and evaluated via experiments in both indoor and outdoor environments.