Robot Localization Using Laser Positioning of Reflectors Based on ICP

Abstract

In this paper, the global localization for a mobile robot in the environment with reflectors arranged in advance is studied. The robot is equipped with a laser that allows to acquire the range and bearing information of the robot with respect to reflectors. After reflector matching, the global pose of the robot can be calculated by localization methods. Focusing on the localization method, a novel laser positioning method based on ICP (Iterative Nearest Point) is proposed, which realizes global localizations by constructing the rigid transformation between reflector sets for two continuous scanning time. In addition, compared with the original ICP algorithm that dealing with point cloud sets, the proposed method has less computation and needs no iteration. The implementation process of the proposed method is derived, and the proposed method is compared with the widely used trilateral localization methods. Experimental results show that our method has higher accuracy and better stability than trilateral positioning method.