Abstract:We consider the median estimation problem in a large-scale sensor augmented RFID system. The large-scale deployment of RFID technology has opened the door to innovative ways to integrate RFID and sensor technology. Sensor-tags are tags that can report over 50 types of physical information to a reader. The traditional way to obtain information from sensor-tags is to query each tag. When the number of tags is large, however, it is prohibitive to query tags individually due to the high delay. In this paper, we present a probabilistic algorithm to estimate the median of a set of sensor-RFID tags without individually querying each tag. The median estimation problem is solved using binary search. Our evaluation demonstrates that the median search algorithm exhibits high accuracy and reasonable time latency. Moreover, we also design an exact algorithm for the continuous median update problem. Our algorithm can incrementally compute the exact median in less time.

Abstract-Radio frequency identification (RFID) tags are used in a lot of applications such as production line and inventory management. Sensors are also a widely adopted wireless technology. This large scale deployment of both RFID technology and sensor technology has opened the door to innovative ways to integrate RFID and sensor technology. Sensor-tags are tags that can report sensed information to an RFID reader. The traditional way to obtain information from tags is to query each tag for its value. This works, however, only for a relatively small number of tags. When the number of tags becomes large it is prohibitive to query one by one due to the high delay. In this paper we present a probabilistic algorithm for estimating the median of a set of RFID tags using binary search. We then evaluate the accuracy and time efficiency of our algorithm.