Multiset Tag Searching Protocol for Multicategory RFID Systems

RFID has been widely deployed in large-scale supply-chain and warehouse scenarios to achieve effective inventory tracking and theft prevention. In such a scenario, there is a practical need for searching for a particular subset with a large number of tags, which can help the retailer to determine whether the current inventory matches the needs of a specific order. The current solution usually searches for the target tags in the entire inventory through single-set matching, which generally takes too much time and is hard to meet given the strict delay requirements of users. To address this problem, this article proposes a technique called multiset tag matching, which applies a hierarchical space-efficient bloom filter variation to aggregate multiple search queries and searches them by category. We show that communication overhead can be optimized further by exploiting the distribution among different categories for grouping similar categories. Thus, each category can be assigned a personalized length of filter to tradeoff between accuracy and cost. Moreover, our results also show that our approach provides a fair degree of accuracy across different categories. We further propose a combination technique for further aggregating category groups with similar sizes to reduce the search and switching costs.