Low Computational Cost Hybrid Approach for Near-Field RFID Tag Localization

In the Ultra High Frequency (UHF) Radio Frequency Identification (RFID) near-field localization context, the two-Dimensional Multiple Signal Classification (2D-MUSIC) technique achieves satisfactory localization performances while requiring a search over all the spatial Direction of Arrival (DOA) and range parameters. This leads to very high computing time cost. On the other hand, Total-Least-Square ESPRIT (TLSESPRIT) originally dedicated to far-field DOA estimation, is less computationally consuming. We propose in this paper to take advantage from the benefits of each of the precedent methods. Our new approach uses far-field TLS-ESPRIT first, in order to obtain coarse estimates of the DOA parameters within a short time. These estimates are then refined thanks to 2D-MUSIC. Several simulations were performed to study the performances of this proposed ESPRIT-MUSIC hybrid algorithm, showing its superiority in terms of accuracy as well as computational time cost. Experiments conducted in an anechoic chamber confirm the efficiency of our approach. The localization results present the same accuracy as near-field 2D-MUSIC method within a shorter time.