IMPLEMENTATION OF RFID TECHNOLOGY FOR OBJECT IDENTIFICATION IN RETAIL: OPTIMIZATION OF AUTOMATED SORTING LINES

Abstract

The problem of integrating UHF radio frequency identification systems into the hardware and software complexes of high-speed conveyor lines in retail warehouses is investigated. Architectural solutions to minimize the impact of multipath radio wave propagation and Doppler frequency shift on the probability of reading passive tags in dynamics are theoretically substantiated. The validation of the proposed hybrid mathematical model of collision processing, combining deterministic prefix tree segmentation (Query Tree) with the stochastic DFSA algorithm of the EPCglobal Gen2 standard, was carried out by developing a comprehensive digital twin in the MATLAB Simulink environment. Simulation modeling proved the architecture's ability to achieve an estimated identification reliability rate of 99.2% at a linear conveyor speed of 3.0 m/s. The architecture of distributed middleware based on the Edge Computing paradigm was virtualized for spatiotemporal filtration of raw radio frequency data and local caching of routing tables. Stochastic simulation results confirm that translating aggregated arrays into the Application Level Events format with subsequent asynchronous transmission via EPCIS interfaces to the Warehouse Management System reduces system latency to 42 ms. The overall estimated throughput of the optimized complex in the simulation environment is 10710 processed units per hour with an algorithmic mis-sort rate of less than 0.08%.