Radio frequency identification (RFID) technology brings tremendous advancements in Industrial Internet-of-Things (IIoT), especially for smart inventory management, as it provides a fast, and low-cost way of counting or positioning items in warehouse. In the last decade, many novel solutions including absolute and relative positioning methods, have been proposed for this application. However, the available methods are quite sensitive to the minor changes in the deployment scenario, including the orientation of the tag and antenna, the materials contained inside the carton, tag distortion, multipath propagation, etc. To this end, we propose a hybrid relative passive RFID localization method (ReLoc) based on both the received signal strength indicator (RSSI) and measured phases, which orders the RFID tags horizontally and vertically. In this paper, phase-based variant maximum likelihood estimation is proposed for lateral positioning, and the RSSI profiles of two tilted antennas are compared with each other for level distinguishing. We implement the proposed positioning system ReLoc with commercial off-theshelf RFID devices. The experiment in a warehouse shows that ReLoc is a powerful solution for practical item-level inventory management. The experimental results show that ReLoc achieves an average lateral and level ordering accuracy of 94.6% and 94.3%, respectively. Notably, when considering liquid or metal materials inside the carton, or tag distortion, ReLoc still performs excellently with more than 93% ordering accuracy both horizontally and vertically, indicating the robustness of the proposed method.Index Terms-Radio frequency identification (RFID), ultra high frequency (UHF), received signal strength indicator (RSSI), phase, multipath propagation, indoor localization, industrial Internet-of-Things (IIoT), inventory management, unmanned aerial vehicle (UAV).
I. INTRODUCTIONS UPPLY chain businesses continue to increase in flexibility and complexity. The ever increasing success of ecommerce requires new supply chain solutions at every stage of operations, including continuous inventory management. Nowadays, most warehouse systems have adopted automatic identification technology such as barcodes or ultra high frequency radio frequency identification (UHF-RFID) tags for automated inventory control, since it helps to minimize the risk of manual errors [1], [2]. However, even though these Manuscript received xx, 2019; revised xx, 2020.