Dynamic Wi-Fi Fingerprinting in Contrastive Embedded Space for Robust Localization
IEEE Sensors Journal · 01.07.2026 · 26 (13) · 19944-19951
DOI: 10.1109/jsen.2026.3690909 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Sensors Journal, erschienen am 01.07.2026. Verfasst von F. Abbasi und 1 weiteren Autor:innen. Die Arbeit nennt 22 Literaturangaben.
Abstract (Originalsprache)
This article presents a novel framework called contrastive learning indoor positioning system (CLIPS) for dynamic indoor positioning using Wi-Fi fingerprinting, which combines contrastive representation learning with sparse modeling and filtering methods. In particular, we use contrastive representation learning to uncover the intrinsic spatial relationships between RSS signals. The method first computes an initial estimate of the location based on the Wi-Fi fingerprint embedding using sparse representation (SR), and then refines the estimate through a local search. In parallel, the Kalman filter handles motion dynamics efficiently, smoothing abrupt transitions and preserving the natural rhythm of spatial evolution over time. The integration of these methods results in more stable and generalizable location estimation. Experimental results show that the proposed method improves localization accuracy compared to traditional methods.
Zitieren
APA
Abbasi, F., & Yazdi, H. S. (2026). Dynamic Wi-Fi Fingerprinting in Contrastive Embedded Space for Robust Localization. IEEE Sensors Journal, 26(13), 19944-19951. https://doi.org/10.1109/jsen.2026.3690909
Vancouver
Abbasi F, Yazdi HS. Dynamic Wi-Fi Fingerprinting in Contrastive Embedded Space for Robust Localization. IEEE Sensors Journal. 2026;26(13):19944-19951. doi:10.1109/jsen.2026.3690909
BibTeX
@article{Abbasi2026,
author = {Abbasi, F. and Yazdi, H. Sadoghi},
title = {Dynamic Wi-Fi Fingerprinting in Contrastive Embedded Space for Robust Localization},
journal = {IEEE Sensors Journal},
year = {2026},
volume = {26},
number = {13},
pages = {19944--19951},
doi = {10.1109/jsen.2026.3690909}
}
EUVIMED