Precise indoor localization of multiple mobile robots with adaptive sensor fusion using odometry and vision data

摘要

An accurate, reliable and a cost effective localization is the key feature of self-navigation of autonomous mobile robots. The position and orientation, together known as pose, of the mobile robot can be determined by using certain localization systems. In this work we use the mobile robot system Robotino - a practice-orientated educational and training system offered by Festo Didactic GmbH. For the Robotinos, position determination can be provided by laser scanner for Robotino or Robotino Northstar System (Festo Didactic GmbH). These existing systems are quite expensive and localization accuracy in certain field dimensions is quite low. In this paper we provide a relatively inexpensive and a more accurate localization system, which combines the strengths of odometry and vision-based localization. The fusion of odometry and vision-based localization data is accomplished with the use of the Extended Kalman Filter (EKF). The resulting localization system delivers better accuracy and more frequent pose information of the mobile robots on the test field.