Self-Localization Based on the doorplate Recognition for Indoor Mobile Robot

摘要

ft is hard for the robot to navigate itself in unknown environments without a map. The robot must know its positions and the destination positions at intervals. Many methods make the robot position itself accurately. However, the robot should be equipped with more precision sensors and move slowly to process the environment information sensors detected. This paper proposes a method for navigating the robot to its destination room by recognizing rooms' number in corridors. We use doorplates, which are attached with a room number respectively, as natural landmarks. The method represents the robot position in the global frame based on the navigation behavior of human. A general architecture of our robot, the strategy of searching doorplate and some image processing methods for the recognition of landmarks and numbers are put forward in this paper. At last, mobile robot CASIA-I we designed and developed is employed to have an experiment in order to testify the validity of the proposed self-localization method.