Localization based on fusion of RFID and stereo image data

A system that localizes and identifies trolleys in the entry area of a distribution center is presented in this paper. The area is monitored by an RFID reading system (consisting of four antennae and a fixed RFID reader) and an inexpensive pair of webcams. The trolleys are equipped with a passive RFID tag and an optical marker. The system uses a Particle Filter to combine Received Signal Strength Indication (RSSI) data from the RFID system and image data from the webcams. We propose procedures to improve RFID and camera localization accuracy. The resulting system achieves an accuracy of less than 30 cm in static and the most likely dynamic scenarios.