Towards low-cost context awareness on smart shelving using passive UHF RFID infrastructure

Abstract

An omni-channel strategy is a method of selling and promoting products that offers customers a comprehensive and cohesive shopping experience. However, this strategy relies on store managers having an accurate, real-time understanding of product availability at all their distribution and retail facilities. Smart shelving is an important avenue for furthering the development of omni-channel retailing and meeting people’s needs. This thesis primarily focuses on the construction of a low-cost context awareness infrastructure for smart shelving using passive UHF RFID tags and radio tomographic imaging (RTI) algorithms.

Firstly, location estimations without fingerprinting in one direction can reach an accuracy of 91.7% on four tested objects. Secondly, the number of stacked layers from 1-3 when placing items on the shelf can be estimated. It is shown that an increase in product volume on the shelf could be related to tag RSSI level changes for five different tested products.

In addition, material classification could be achieved by tag RSSI attenuations. Tests are done between three classes (metal, glass, and plastic), with three objects each class. In the three-location tests, it is possible to clearly differentiate between three types of materials based on the value of variations in tag RSSI attenuations.

Finally, the integration of battery-free environmental sensors is accomplished by incorporating an RFID tag equipped with resistance measurement capability and a photoresistor. By measuring the resistance of the photoresistor, the designed light sensor could provide additional information (besides the tag RSSI change) about the volume of material on a shelf. Moreover, this can be done using only a single UHF RFID Gen 2 protocol.