Study On Cleaning Technology For RFID Data Based On Time Series In Internet Of Things

The high quality data can provide important basis and the wizard to decision-making for the enterprise. With the development of the Internet of Things and RFID technology, domestic and foreign scholars take wide attention to the RFID uncertain data cleaning technology. At present, RFID uncertain data cleaning technology has achieved initial results, but the studies don’t make full mining of the time relevance among data.On the basis of a detailed introduction about strengths and weaknesses of the proposed frameworks and algorithms, the RFID data cleaning framework based on selection mechanism is established. According to the mass and uncertain of the data, the data entered into the framework will be scanned and detected, and the framework determines whether the data should be transmitted to the processing layer. In the framework, the data only choose the best cleaning route but not need to travel all the layers in order to reduce the cleaning time and improve the efficiency.Combining with the time-series characteristics, some cleaning algorithms aimed at the different kinds of uncertain data are proposed on the basis of the CFBS. The cleaning algorithms are based on the time series and can clean the uncertain data hierarchically .At the same time, the algorithms introduce the concept of tag response rate which considers the response times and the response time to clean uncertain data.Finally, the experiments between the methods proposed in this paper and the traditional cleaning methods are analyzed. The experimental results show that the CFBS cleaning framework has a good time performance and can reduce the data cleaning time consumption effectively, the efficiency of cleaning improves obviously; the cleaning algorithms based on time series are not only good at the static tags cleaning, but also can detect the dynamic tags real-time.