ID filtering and format translation become imminent performance bottlenecks of RFID middleware in anticipated massive traffic in the automated RFID systems. An RFID middleware manages RFID Readers/Writers, translate the IDs in received events, filter them by registered subscriptions, and notify the aggregated reports. We redesign the data flow in the middleware with byte-based filtering engines that selectively adjust to dynamic situations. Our Open-Source implementation of the proposed methods demonstrated the significant performance boost from the legacy design by reducing the latency from more than 1 ms to under 100us to process 1000 events on a commodity PC.
In order to prevent food safety and counterfeit, it is effective to assign serial numbers to each item and to manage their information for each product with a computer. The JAN barcode currently in use is for product identification and was not suitable for unique identification. The JAN barcode has fine print irregularity, and their shapes are different even with barcode of the same beverage. In this research, we have developed a technology to identify individual JAN barcode by detecting fine print irregularity of barcode, feature quantity extraction, and encoding with the technology of image processing / image recognition, and establish a technology to simultaneously realize product identification and individual unique identification with only JAN barcode.
In source tagging era, we can uniquely identify each merchandise with an ID encoded in barcode or RFID tag. The source tagging facilitates and produces new services such as sharing wardrobes with friends, valuation of used things according to the trade and usage history, food consumption based on their expiration dates and automatic notification of product recall. In this exhibition, we demonstrate a prototype of information system in the source tagging era.