Research On Clean Technology Of RFID Sensing Data Of Manu-Facturing In IOT

RFID technology as the core technology in IOT is widely used in many industries, as information technology continues to depth, the data becomes the most critical element in information system, many of the current large-scale RFID data is an important source of data systems, faced with a huge weak semantic data from the physical world, how to effectively extract the raw data is crucial.Pipe production information traceability process will produce tremendous raw tag data, if the data directly into the top decision-making system that will bring huge amount of computation to upper application or even impossible to get accurate decision results, significant losses to the enterprise, so cleaning RFID raw data is very important. The main source of dirty RFID data is leakage read, Read more, redundancy, etc., data leakage as the main source of dirty data in Sparse topology and data redundancy as the most important source of dirty data in intensive topology. For this, read this article from the RFID data leakage and data redundancy are two ways to describe, and discuss practical RFID application system efficiency of different algorithms.This paper studies the optimization of RFID read data leakage and redundancy algorithm, the main work:(1) Comprehensive analysis of the characteristics of RFID system components and data flow, analysis commonly used cleaning strategies in RFID network and highlights a few classic RFID data cleaning algorithms.(2) To meet the data integrity of RFID network, and improve reading data leakage to fill the correct rate, on the basis of research on algorithms SMURF, Propose an adaptive window size adjustment algorithms ADCLM. The algorithm combines statistical sampling theory, make decisions based on data integrity and dynamic testing conditions, also for the treatment of marginalization on the label, large window stain effect, small window shake effects and other issues to do a more in-depth treatment.(3) To minimize redundant data in the RFID data network, In the study, based on the algorithm RRE and LEO, propose a specific topology does not depend on a specific reading order based middleware to redundancy reader hybrid algorithm. The algorithm ensures data integrity in meeting the requirements of the entire RFID application system activation reader at least. Finally, in the Piles RFID traceability management system, we do the experimental test algorithm performance, experimental results show that smoothing algorithm of leakage and the redundancy algorithm can meet the real-time processing of massive data streams, can get the complete upper application available for use semantic data.