Localization using relative mapping technique for mobile soccer robots

Abstract

Localization is an important topic in the field of mobile robotics. It is the process of estimating the position and orientation of a mobile robot. Several probabilistic techniques are available for this purpose, some of which use a compass sensor to directly determine current orientation. This sensor is highly susceptible to magnetic interference. Here, we propose an association based approach which solely relies on camera images to localize in a pre-defined environment. We use the Monte Carlo Localization for position based filtering. Further, we present a relative landmark mapping technique and white points based filtering used to obtain the final pose of the robot after incorporating its orientation. To check the accuracy of the proposed methodology, it was tested on a large number of realworld test cases. The technique resulted in a high success rate and accurately estimated the position and orientation of the mobile robot in the field. It is currently being employed for localizing humanoid robot, AcYut, in humanoid robot soccer games.