On Information Collection in Multi-Tagged COTS RFID Systems

We study the problem of target object information collection in multi-tagged COTS RFID systems. Unlike its singletagged peers, the multi-tagged COTS RFID scenario poses new challenges in devising information collection algorithms: 1) Tags attached to the same object carry identical information. Hence, reusing single-tagged information collection algorithms leads to unnecessary redundancy; 2) Multi-tagged RFID systems are often deployed in applications where tags are vulnerable to damage. Such faulty tags may severely degrade the performance of information collection; 3) Most state-of-the-art information collection algorithms rely heavily on the hashing operation that is not seamlessly supported by the C1G2 standard, rendering these solutions inefficient and impractical, especially in largescale RFID systems. To tackle these technical challenges, this paper makes three contributions. First, we develop an efficient and compact tag pseudo-ID design, enabling the reader to select a single tag from each target object to collect information with only one SELECT command. Second, we construct a robust faulthandling mechanism capable of recognizing faulty tags without executing the entire slot. Third, armed with the above two techniques, we develop a novel information collection algorithm by leveraging the functionality offered by C1G2 to optimize the information collection sequence, thus minimizing the overall execution time. Empirical experiments on a COTS RFID system prototype demonstrate that our algorithm outperforms the best existing solution by 35-50% on average.