An Improved VSLAM for Mobile Robot Localization in Corridor Environment

Localization is a fundamental capability for an autonomous mobile robot, especially in the navigation process. The commonly used laser-based simultaneous localization and mapping (SLAM) method can build grid map of indoor environment realize task. However, when robot comes to long corridor where there exists many geometrically symmetrical similar structures, it often fails position itself. Besides, not represented semantic level that cannot interact well with users. To solve these crucial issues, this paper, we propose improved visual SLAM approach robust precise global localization. system divided into two main steps. first step construct topological using SLAM, text detection recognition, laser sensor data. second which repeats part work but makes best use prebuilt map. Experiments show our solutions perform localize successfully almost everywhere while traditional methods fail.