Error Recovery of Autonomous Mobile Robot With Global Matching

Abstract

This paper proposes a new navigation system for an autonomous mobile robot. An earlier version of the robot navigated by recognizing a single landmark at a time, but the landmark was often lost sight of, and as a result, the robot's self-position was lost. We propose a method that enables the robot to recover from such an error. The method determines the self-position by using information on the corner, windows and doors of a room simultaneously if the single landmark is lost.