Research On Fast Detection Algorithm For Label In Large Scale Active RFID System

Radio frequency identification (Radio Frequency Identification, RFID) is an automaticidentification technology developed in the eighty’s, and an important part of the Internet ofthings technology, that can be widely used in every fields. Now it has been widely used inmonitoring and recognition, intelligent transportation, logistics management, industrial, medical,security etc. In recent years, with the development of the Internet of things technology andapplication, RFID identification technology as the key technology of the Internet of things hasbecome a hot research at home and abroad.RFID technology has the characteristics of non-contact. It can quickly identify thehigh-speed moving objects and auto-maticly identify multiple remote tags. The active tag ofRFID with its own battery has the advantages of remote information communication support,large amount of data transmission and less emissive power, so it has a very broad applicationareas, particularly used in the monitoring and recognition of large valuables or dangerous goods.Minimizing the recognition time, reducing the energy consumption of the active tag andimproving the detection efficiency of the existence of label is an hots issue of loss label detectionalgorithm in the large system composed of active RFID.This paper briefly introduces the knowledge of the research background and significance ofthis topic, RFID technology system principle of work and other related. This paper makes athorough research in view of the existence question of the label in order to improve theefficiency of active tag label inspection and to reduce power consumption of the active tag andimproves mainly from two aspects: the problem of reader work in parallel conflict in multi-reader system and the problem of energy consumption and efficiency in the process of detection.There is cross collision problem of reading area when some readers work in parallel and tosolve the reader collision problem is always a research hots of RFID Technology. This papermakes improvements based on the distributed anti collision algorithm combining particle swarmalgorithm(PSO) and graph coloring algorithm. The central idea is the server optimizes andupdates the reader slot allocation in real time according to the reader’s tag number estimation.New algorithm uses the slot of the reader network maximum, reduce the slot waste of the readerand ultimately improves the overall efficiency of system identification.Tag identification technology is the key technology of RFID in RFID system. Automaticidentification technology of traditional determines the label existence problem by collecting allthe tags identification number and comparing the tags identification number to the tag set storedin the database. But the identify efficiency of this kind of method is very low and its power consumption for active tags is higher. In view of the above question, this proposes a new tagdetection algorithm of HPL in the multi-reader parallel basis which is improved based on frameslotted protocol. The main content of this algorithm is the reader launch query frame carrying arandom number Q every time and the label selects a slot through the selection of Q value toreplies a bit. The reader completes a recognition process through a fixed period polling and forma matrix according to the reply information to estimate the existence of label. Theoreticalanalysis and simulation prove that Reader scheduling mode used in this paper can avoid thereader collision collision problem, improve system throughput and reduce recognition time.Compared with the traditional detection algorithm based on label number comparison, the HPL(Hash polling Detection) algorithm proposed in this paper can reduce one order of magnitude inthe tag energy consumption, and also shorten the execution time of the algorithm. Comparedwith the current best iterative detection algorithm for ID-free protocol, the HPL (Hash pollingDetection) algorithm can reduce three orders of magnitude in the tag energy consumption, andalso shorten the execution time of the algorithm.