Filtering Redundant Data from RFID Data Streams

Radio Frequency Identification (RFID) enabled systems are evolving in many applications that need to know the physical location of objects such as supply chain management. Naturally, RFID systems create large volumes of duplicate data. As the duplicate data wastes communication, processing, and storage resources as well as delaying decision-making, filtering duplicate data from RFID data stream is an important and challenging problem. Existing Bloom Filter-based approaches for filtering duplicate RFID data streams are complex and slow as they use multiple hash functions. In this paper, we propose an approach for filtering duplicate data from RFID data streams. The proposed approach is based on modified Bloom Filter and uses only a single hash function. We performed extensive empirical study of the proposed approach and compared it against the Bloom Filter, d-Left Time Bloom Filter, and the Count Bloom Filter approaches. The results show that the proposed approach outperforms the baseline approaches in terms of false positive rate, execution time, and true positive rate.