This paper presents an experimentally validated localization method for underground targets based on low-frequency alternating magnetic fields generated by surface anchors and measured by an underground receiving coil system. The proposed approach uses time-sequential excitation of the transmitting coils (i.e., surface anchors), extraction of frequency-domain features from the received signals (underground captured by a tri-axis receiving coil), and a model-based estimator to infer the target position. The full localization pipeline is tested in a real environment to assess its robustness under non-ideal hardware behavior and measurement noise. Experimental results show that the positioning RMSE remains below 25 cm for the x, y, and z coordinates, with a median error of 12 cm or less. For the overall position error, the RMSE is 35 cm, and the median is 30 cm. In addition, the errors remain below 35 cm for about 90% of the tested tag positions for x, y, and z coordinates, while the total positioning error remains below 55 cm with the same probability level. These results indicate that the proposed magnetic localization framework is a promising solution for through-soil localization, as an example for dronebased autonomous drilling and mapping applications.
Real-world Validation of AC Magnetic Localization System for Underground Targets
Provenzale, C.;Milano, F.;Capriglione, D.;
2026-01-01
Abstract
This paper presents an experimentally validated localization method for underground targets based on low-frequency alternating magnetic fields generated by surface anchors and measured by an underground receiving coil system. The proposed approach uses time-sequential excitation of the transmitting coils (i.e., surface anchors), extraction of frequency-domain features from the received signals (underground captured by a tri-axis receiving coil), and a model-based estimator to infer the target position. The full localization pipeline is tested in a real environment to assess its robustness under non-ideal hardware behavior and measurement noise. Experimental results show that the positioning RMSE remains below 25 cm for the x, y, and z coordinates, with a median error of 12 cm or less. For the overall position error, the RMSE is 35 cm, and the median is 30 cm. In addition, the errors remain below 35 cm for about 90% of the tested tag positions for x, y, and z coordinates, while the total positioning error remains below 55 cm with the same probability level. These results indicate that the proposed magnetic localization framework is a promising solution for through-soil localization, as an example for dronebased autonomous drilling and mapping applications.| File | Dimensione | Formato | |
|---|---|---|---|
|
2026_Conference_MNC_Localization_Underground.pdf
solo utenti autorizzati
Tipologia:
Documento in Pre-print
Licenza:
Creative commons
Dimensione
3.75 MB
Formato
Adobe PDF
|
3.75 MB | Adobe PDF | Visualizza/Apri Richiedi una copia |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

