Mobile robot localization using the phase of passive UHF-RFID signals

This paper presents a Global Localization system for an indoor autonomous vehicle equipped with odometry sensors and an RFID (Radio Frequency IDentification) reader to interrogate tags located on the ceiling of the environment. The RFID reader can measure the phase of the signals coming from responding tags. This phase has a non univocal dependence on the distance robot-tag but, in the considered frequency, it is really sensitive to a change in the position of the robot. For this reason, a Multi-Hypothesis Kalman filtering approach provides a really satisfactory performance even in the case a very small density of tags is used: in the experimental tests, an average position estimation error of about 4 cm is achieved using only two tags for an area of about 5 square meter.