2D Robot Localization with Image-Based Panoramic Models Using Vertical Line Features

This paper presents a method for estimating the position of a mobile robot in an indoor environment. The proposed technique uses a model of the environment formed by two panoramic cylindrical images taken at di erent locations, and a planar image taken at the current position. The current position and orientation of the robot are then computed without any additional information. We assume that the robot is moving on a plane ( oor plane), which is very common for indoor environments. We will extend our previous result on using point features and exploit vertical line features to perform the localization. Our method is attractive because it does not require an explicit 3D model of the environment, and the location of the camera is not restricted to positions close to the cylindrical models. We will describe experimental results for both synthetic and real data to demonstrate the e ectiveness of the proposed method.