Exploiting id knowledge for tag identification in rfid networks

In this paper, we consider the problem of identifying a set of objects in an RFID network. We propose a query tree based protocol to reduce the number of transmission collisions. The key idea of our protocol is to take advantage of the (partial) knowledge about the set of tags to be identified: in particular, their approximated number, and above all their distribution. More specifically, we show how much gain in performance can be achieved by having information about the set of IDs to be identified. Such information can be obtained from statistical data, or from other features, depending on the application. Of course, a perfect knowledge makes useless the identification process: the aim of this work is to show how query tree protocols improve, by taking into account of the information about the tags. Simulation results show that our approach performs better than classical query tree protocols [1], in terms of number of queries needed to identify all tags, which is a commonly used metric, strictly related to delay. Moreover, our approach out performs also Aloha-based protocols that for their randomized nature, are independent from the IDs distributions: in such protocols, tags randomly choose a slot in which transmit, that is equivalent to "re-naming" the IDs, so they do not take any advantage from the knowledge of the original ID distribution.