Robot Localization Algorithm using Odometry and RFID Technology

This paper presents a localization algorithm based on the particle filter using odometry and RFID sensor readings. The integration of both measurements permits to localize a mobile robot with a bounded error estimation, while dead reckoning approaches using only odometry let the error diverge. RFID tags are integrated in the environment; it is supposed their positions have been learnt during a learning step not considered here. The localization algorithm is based on the observation model for the RFID measurements; this model is determined experimentally. The algorithm and the RFID reader are embedded and evaluated on a mobile robot; experimental results show that our localization algorithm gives the robot position with a bounded error when using RFID measurements, additionally in order to correct odometry drifts.