RFID Redundancy Data Filtering Research Based On Sliding Windows

Filtering redundant data is very important for radio frequency identification(RFID) applications, whose purpose is to remove the large amount of redundant data from RFID data streams and to provide accurate and effective data for upper applications. Due to the characteristics of RFID systems, the tagged object usually stays in the detection range of a reader or moves slowly, so the reader may generate a lot of duplicate data during the process of detection. The duplicate data doesn’t provide any valuable information to the upper applications, but they will take up system resources and affect system efficiency. Therefore, before the RFID data are sent to the applications, how to use the limited memory space in the limited time to remove these redundant data is one of the key problems to be solved in RFID data streams.This paper first introduces the RFID system model and the characteristics of RFID data streams, and figures out the way that duplicate removal of RFID data streams is different with the duplicate removal of traditional data and data streams. Then, the related duplicate removal techniques, including traditional data, bloom filter, data stream window model and data streams, are introduced in details, and their advantages and problems are also analyzed.For the case of independent reader detection ranges, in order to make the RFID system identify location change timely and filter the time redundancy data effectively, a Temporal Spatial Bloom Filter (TSBF) for sliding window model is proposed. Based on the standard Bloom filters, TSBF uses a two-dimensional array not only to store reader IDs and timestamps for identifying both time redundancy data and location movement, but also to provide real-time and accurate information for the upper applications. Meanwhile, in order to ensure that the false positive rate does not increase for the reason that the filter becomes full, a decay removal algorithm for the expired elements is proposed, which only introduces extremely low false negative rate, so that it ensures the false positive rate will not increase because of large amount of data. Through experiment comparative analysis, TSBF algorithm can effectively filter time redundancy data and has a good performance in the case of location movement.To deal with the large-scale RFID environment with cross detection range and the serious influence of massive data on network transmission load, we propose a two phase filtering algorithm combining local filtering and global filtering. In the local filtering, we use a Max Counting Bloom Filter (MCBF) to remove the time redundancy data generated from readers. In the global filtering, a TSBF-S algorithm is proposed to remove the space redundancy data. The algorithm simplifies the judgment condition of TSBF so that it can deal with space redundancy data, and can improve execution efficiency in the premise of ensuring accuracy rate. Through experiment analysis, the operating efficiency of two phase filtering algorithm is much higher than that of TSBF algorithm.