Advanced Distance Estimation in ESP32‐S2 IoT Devices Using Machine Learning‐Enhanced FTM
International Journal of Communication Systems · 10.09.2026 · 39 (13)
DOI: 10.1002/dac.70573 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Communication Systems, online erschienen am 22.07.2026. Verfasst von Limin Hou. Die Arbeit nennt 21 Literaturangaben.
Abstract (Originalsprache)
Accurate distance estimation based on Wi‐Fi Fine Time Measurement (FTM) round‐trip time (RTT) observations is essential for practical Internet of things (IoT) localization systems. This study investigates a machine learning–based error‐correction framework for improving FTM ranging accuracy using low‐cost ESP32‐S2 devices. A public ESP32‐S2 FTM dataset was employed in terms of indoor, outdoor and semianechoic environments, and four regression models were assessed: gradient boosting, random forest, decision tree, and XGBoost. The correction term was the models trained on the difference between the ground‐truth distance and the distance value estimated from the physics‐based RTT model. Using the full pipeline, including feature engineering and isotonic calibration, the best overall configuration reduced the mean absolute error (MAE) from 3.41 m for the physics baseline to 0.97 m. Environment‐specific analysis showed that random forest achieved submeter MAE in outdoor scenarios (0.071 m at 20 m and 0.175 m at 40 m), while indoor MAE was approximately 0.56–0.61 m across models. Performance in the semianechoic environment remained substantially weaker, with MAE above 1.4 m and R 2 (coefficient of determination) roughly 0.32. The main contribution lies in an integrated seven‐stage ranging pipeline that combines physics‐based initialization, error‐target learning, engineered features, model tuning, output calibration, and environment‐specific interpretation. The results show that postprocessing the measurement using software can significantly enhance the accuracy of FTM ranging on the ESP32‐S2, while leaving the measurement hardware unchanged.
Zitieren
APA
Hou, L. (2026). Advanced Distance Estimation in ESP32‐S2 IoT Devices Using Machine Learning‐Enhanced FTM. International Journal of Communication Systems, 39(13). https://doi.org/10.1002/dac.70573
Vancouver
Hou L. Advanced Distance Estimation in ESP32‐S2 IoT Devices Using Machine Learning‐Enhanced FTM. International Journal of Communication Systems. 2026;39(13). doi:10.1002/dac.70573
BibTeX
@article{Hou2026,
author = {Hou, Limin},
title = {Advanced Distance Estimation in ESP32‐S2 IoT Devices Using Machine Learning‐Enhanced FTM},
journal = {International Journal of Communication Systems},
year = {2026},
volume = {39},
number = {13},
doi = {10.1002/dac.70573}
}
EUVIMED