IoT Anomaly Traffic Detection Method Based on Transformer Encoder Architecture
Journal of Computer Science and Frontier Technologies · 30.05.2026 · 3 (2) · 177
DOI: 10.63313/jcsft.9081 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Computer Science and Frontier Technologies, online erschienen am 30.05.2026. Verfasst von Yuxuan Jiang. Die Arbeit nennt 11 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-SA 4.0).
Abstract (Originalsprache)
The widespread deployment of IoT devices has made heterogeneous network traffic highly complex, making abnormal traffic detection vital for IoT security. This paper presents a Transformer-based IoT abnormal traffic detection method tailored to the long-range temporal correlation of IoT traffic and the drawbacks of existing approaches. It first builds traffic embeddings with learnable positional encoding to project raw traffic features into a high-dimensional space, then leverages Transformer multi-head self-attention to capture long-range dependencies by modeling temporal correlations in traffic sequences. Feature compression and residual enhancement modules are further adopted to lower model complexity and inference delay for resource-limited IoT terminals. Experiments verify that the scheme accurately identifies normal traffic and diverse anomalous attacks with outstanding detection accuracy and stability.
Zitieren
APA
Jiang, Y. (2026). IoT Anomaly Traffic Detection Method Based on Transformer Encoder Architecture. Journal of Computer Science and Frontier Technologies, 3(2), 177. https://doi.org/10.63313/jcsft.9081
Vancouver
Jiang Y. IoT Anomaly Traffic Detection Method Based on Transformer Encoder Architecture. Journal of Computer Science and Frontier Technologies. 2026;3(2):177. doi:10.63313/jcsft.9081
BibTeX
@article{Jiang2026,
author = {Jiang, Yuxuan},
title = {IoT Anomaly Traffic Detection Method Based on Transformer Encoder Architecture},
journal = {Journal of Computer Science and Frontier Technologies},
year = {2026},
volume = {3},
number = {2},
pages = {177},
doi = {10.63313/jcsft.9081}
}
EUVIMED