Research On Reliability Of Data Reading In The Large-scale RFID Network

In an era of rapid technological progress, the Internet of Things (IoT) is no longer anovel concept, and with its development, Radio Frequency Identification (RFID) is widelyused in many respects of logistics, localization, food traceability, etc. RFID is a non-contactwireless communication system that uses radio frequency signal to transfer information datafrom a tag attached on an object to a reader so as to complete automatic identification ofobjects or traceable applications. As the RFID technology developed, its relevant technicalstandards have refined. Thereinto, ECP global C1G2, one of the most popular standards, iswidely approved in the860MHz to960MHz band.RFID, which has the fast multi-objects identification capacity as its major characteristic,is subject to wireless channel frequency and timing similar to other various wirelesscommunication technologies, and the RFID Collision Problem (RCP) exists commonly in thesystem of multi-readers or multi-tags. In the large-scale RFID network, the existence of signalcollision is a great challenge to the reliability of data reading. Past researches have showedthat a stable, efficient and distributed processing anti-collision algorithm plays an importantrole for improving the reliability of data read in the large-scale RFID networkDepending on the different application environment, the solution to RCP could bedivided into the tag anti-collision algorithm and the reader anti-collision algorithm. RFID taganti-collision algorithm mainly consists of the PA (Pure ALOHA) algorithm, SA (SlottedALOHA) algorithm, FSA (Frame Slotted ALOHA) algorithm based on ALOHA, SBT(Simple Binary Tree) algorithm, BS (Binary Search) algorithm, QT (Query Tree) algorithm,CT (Collision Tree) algorithm based on binary tree structure and several improved algorithms.RFID reader anti-collision algorithm could be divided into power control and central controlrespectively. The classical reader anti-collision algorithm which is based on central controlconsists of Colorwave algorithm, HiQ algorithm, Pulse algorithm, etc. Through a lot morelearning and analysis on RFID reader anti-collision algorithms, profound understanding of thedata read reliability of the large-scale RFID network is formed, and based on that, improvedresearch of Colorwave algorithm is proposed. Based on the classical reader anti-collision algorithm Colorwave, a graph Coloring andNon-completely Random back-off based Colorwave reader anti-collision algorithm (CNRC)is proposed. The main innovations are as follows:1. The determination method of in-round chromatic number initial value in the CNRCalgorithm based on the graph coloring theory is proposed. In its infancy, determine theexecutive initial value and act as one of the algorithm inputs on the basis of graph coloringtheory.2. The non-complete random back-off mechanism of reader in the collision stage basedon the comparison of randomly generated numbers is proposed. Once any collision occursafter each transfer attempt, the readers generate random numbers and compare each otherinstantly. The reader which had the minimal random number does not change its color, whilethe others perform random back-off mechanism normally.3. The control channel is added as the special forwarding channel for Kick commandsbetween readers. Besides the data channel in the Colorwave, CNRC adds a control channeldedicated to the Kick commands transmission between readers.Under different reader density environment, the experiment completes simulationanalysis on changes of the average system throughput and delay ratio over the system runningtime and reading request occurrence. The simulation results demonstrate that in dense readerenvironment, CNRC has better performances compared with the traditional readeranti-collision algorithm.