Moc-SAR: Motion Compensation for Synthetic Aperture-Based Holographic Localization in UHF RFID System

With the development of the industry's digital transformation driven by 5G-Advanced, an increasing need for higher precision arises in using radio frequency identification (RFID) to enable product traceability across the production chain, warehouse management, and the retail network. The past decade has witnessed a significant advancement in phase-based methods for object localization, with a growing emphasis on accuracy. However, deviations in phase values caused by signal attenuation, Doppler frequency shift, and measurement-induced error during the movement process are typically disregarded. In this article, Moc-SAR, a motion compensation for synthetic aperture-based holographic localization method via a moving UHF RFID system, is proposed. In this framework, a scheme of data processing based on synthetic aperture measurement, removal of phase jumps and stationary samplings, phase differential, and data grouping is proposed to construct the dataset. Then the XGBoost-based motion compensation model is trained considering potential influenced factors such as phase, Doppler frequency, received signal strength (RSS), and coordinates of virtual antenna elements. The prediction results of phase are used for holographic localization. The variations in the dataset at each stage during the scheme of data processing are visualized. Statistical analysis of root mean square error (RMSE) illustrates that the trained XGBoost model achieves satisfactory predictive performance. Further numerical analysis of the phase offset factors is conducted based on the measured data. Experimental results of the test dataset used for holographic localization show that Moc-SAR performs superiorly on accuracy, effectively dealing with deviation on phase measurements in multipath environment.