Wireless User ID Identification Based on Incremental Learning
IEEE Communications Letters · 2026 · 30 · 1380-1384
DOI: 10.1109/lcomm.2026.3673918 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Communications Letters, 2026. Verfasst von Yiyun Zhang und 3 weiteren Autor:innen. Die Arbeit nennt 16 Literaturangaben. Gefördert von: National Science and Technology Major Projects of China, Beijing Municipal Natural Science Foundation, Natural Science Foundation of China.
Abstract (Originalsprache)
With the development of wireless sensing technology, WiFi-based identity recognition has gained widespread attention in fields such as human-computer interaction and home security. Although WiFi-based sensing has seen initial research, existing identity recognition methods primarily rely on fixed user categories and lack adaptability to dynamic identity changes. Furthermore, the limited number of samples for new identity classes poses a significant challenge. In this letter, we propose a WiFi-based identity recognition method based on few-shot incremental learning, which enables effective adaptation to scenarios with small sample sizes and evolving user identities. We introduce a new few-shot training strategy and a continuously evolving classifier during the user training phase to avoid severe overfitting and catastrophic forgetting. Moreover, we conduct comprehensive comparative experiments on our method and other systems in scenarios with fixed user identity categories. The experimental results show that the proposed method effectively mitigates performance degradation across incremental sessions in terms of accuracy, precision, recall, and F1-Score. In scenarios with an increasing number of identity categories, the proposed method maintains an accuracy of 84% despite limited sample availability, consistently outperforming baseline methods.
Förderung
- National Science and Technology Major Projects of China
- Beijing Municipal Natural Science Foundation
- Natural Science Foundation of China
Zitieren
APA
Zhang, Y., Wang, G., Gao, F., & Gong, W. (2026). Wireless User ID Identification Based on Incremental Learning. IEEE Communications Letters, 30, 1380-1384. https://doi.org/10.1109/lcomm.2026.3673918
Vancouver
Zhang Y, Wang G, Gao F, Gong W. Wireless User ID Identification Based on Incremental Learning. IEEE Communications Letters. 2026;30:1380-1384. doi:10.1109/lcomm.2026.3673918
BibTeX
@article{Zhang2026,
author = {Zhang, Yiyun and Wang, Gongpu and Gao, Feifei and Gong, Wei},
title = {Wireless User ID Identification Based on Incremental Learning},
journal = {IEEE Communications Letters},
year = {2026},
volume = {30},
pages = {1380--1384},
doi = {10.1109/lcomm.2026.3673918}
}
EUVIMED