Indoor SLAM using a range-augmented omnidirectional vision

Abstract

A promising combination of sensors as a laser range finder and an omnidirectional camera can be used to extract rich 3D information from indoor environments for robot mapping and localization. This paper presents an implementation of FastSLAM using 3D vertical edges retrieved from a range-augmented omnidirectional vision sensor. Our sensor model in conjunction with the FastSLAM algorithm solves the indoor Simultaneous Localization and Mapping problem. The real world experiment to validate our approach used a Pioneer-3DX mobile robot equipped with a URG-04LX laser range finder and an omnidirectional camera.