Anti-collision algorithm for RFID system using adaptive Bayesian Belief Networks and it's VLSI Implementation

Abstract

In Radio Frequency Identification system, reader collision problems are generally mitigated by maximizing the total effective interrogation area of an RFID reader network or by automatic adjustable frame size of reader, etc. This paper proposes a novel anti-collision algorithm for RFID system using adaptive Bayesian Belief Networks. Bayesian belief networks are suitable for target recognition within a network using the category, identity and type or class of a target. These three attributes are being modelled based on the theory, which relies on previous and related trials approach. It is a classical problem which uses the observers' belief that an event will occur or not. Similarly, it can predict from previous trial reports whether a node is reliable or not. An RFID network faced with the reader collision problem when multiple readers are deployed densely and enters the region of another reader. Tags receive signal from more than one reader and instant identification by a single reader becomes a problem. The aim of this paper is to develop an efficient anti-collision algorithm, where using BBN, the reader identifies foreign elements like unwanted tags and readers and detect the specific/authentic tags. Similarly, the tag can respond the specific reader only and rejecting signal from unwanted readers, thus control the transmitting characteristics and minimizes the collision problems. In this paper, the authenticity checker is developed and implemented for RFID system using VHDL code and simulated the design with Xilinx ISE 14.3 simulation tools and high performance FPGA board.