Global Localization for the Mobile Robot Based on Natural Number Recognition in Corridor Environment

This paper proposes global localization for mobile robot by introducing local goal based navigation and model-based object recognition. In navigation stage, the robot follows the wall while detecting a door using a laser scanner, and then sets up the local goal near detected door. In recognition stage, room number is recognized and also ambiguous room number is rejected by multistage rejection method (MRSM) in order to reduce false recognition. Recognition results by various methods are demonstrated and room number feature map is built after exploring the whole corridor of LG research center.