Research On Data Processing Method For RFID Middleware

RFID(Radio Frequency Identification)is widely used in object tracking,intelligent storage and anti-counterfeiting because of its characteristics of non-contact and low cost.One of the reasons preventing it from spreading more widely is that users may face many problems when attaining RFID data,such as redundancy and missing of data.In this paper,the background and current situations of RFID and middleware of RFID are introduced in detail firstly.Then,two questions about middleware's module of data processing are fully analyzed: how to deal with redundancy and how to fill up the missing of data.Following aspects are discussed in this paper:(1)Principle of RFID is analyzed.The units of system as well as functions,struction and technical standards of RFID middleware and the characteristics of data stream are summarized.(2)To deal with redundancy of data,redundant readers are firstly studied in detail.The new algorithm based on middleware called MCBA is proposed after the analysis of RRE,LEO and CBA.The improved algorithm combined with heuristic strategy has advantages on calculating weight to select readers more reasonably and do not need to write in tags.Experimental results show that MCBA can improve accuracy of finding redundant readers and reduce energy consumption.(3)For the case of redundancy of data-level,advantages and disadvantages of existing algorithm are summarized to let the EPC information of tags distinct as possible.The improved algorithm based on SNM is combined with the characteristics of RFID data stream and make the sorting twice on timestamp and reader ID.The size of window is also self-adapting during duplicates removing to decrease matching time by means of recording times of repetition.Experimental results show that the algorithm can remove more repetitions.(4)SMURF is further studied due to filling up data missing.It is weak in detecting the state of tag and adjustment of size of window.The new detecting mechanism and the strategy of adjusting size of window are designed.Markov chain is used in new algorithm to detect the state of tag in accordance with state-transition-probability matrix calculated by observed values.Probability is used in adjusting size of window to make the adjustment more reasonable.MC-SMURF and optimizations is proposed.Experimental results show that the new algorithm works better.