Skip to Main Content
In this paper, we present a novel radio-frequency identification (RFID) smart shelf that accurately locates tagged objects using standard passive UHF RFID tags. This standard-based commercial off-the-shelf approach provides significant advantages over custom HF RFID and other near-field RFID approaches, including reduced tag costs, minimal infrastructure costs, and simple operation. In order to achieve accurate location sensing of objects sitting on the shelf, we utilize a novel localization algorithm that utilizes detected changes in a tag's readability to infer the presence of neighboring tags. According to our experimentation results, with a single RFID reader antenna for two wooden shelves of size 91 cm × 152 cm, our smart-shelf system estimates nine box-level object locations with an average error of just 18.48 cm, which is a 71% improvement in accuracy compared with the previously published k nearest neighbor (KNN) algorithm.