CEEMDAN ‐ AGNN ‐ GTCN : A Denoising‐Enhanced Spatiotemporal Graph Neural Network for IIoT ‐Based EV Charging Demand Forecasting on Icy and Snowy Roads
Internet Technology Letters · 2026-07 · 9 (4)
DOI: 10.1002/itl2.70267 ↗Kurz gefasst
Fachartikel in der Zeitschrift Internet Technology Letters, online erschienen am 15.05.2026. Verfasst von Xiaoming Lu und 5 weiteren Autor:innen. Die Arbeit nennt 12 Literaturangaben.
Abstract (Originalsprache)
Accurate prediction of electric vehicle (EV) charging demand on icy/snowy roads is challenging due to nonstationary, noisy consumption patterns, which conventional models struggle with in Industrial Internet of Things (IIoT) environments. This study develops a robust spatiotemporal framework leveraging multisource IIoT data to capture these complex characteristics, supporting winter traffic‐energy coordination and resilient infrastructure. METHODS: We propose CEEMDAN‐AGNN‐GTCN, a hybrid IIoT‐enabled model processing raw industrial data. It applies Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) and wavelet thresholding for noise reduction. An Attention‐enhanced Graph Neural Network (AGNN) captures spatial dependencies from the IIoT network topology, while a Gated Temporal Convolutional Network (GTCN) extracts multiscale temporal features. Using real‐world icy road IIoT data, the model achieved optimal performance with a 2‐h sliding window (MAPE ≈4.82%, R 2 ≈0.94) and demonstrated robustness, with only a 9.6% RMSE increase for 30‐min ahead forecasts. The framework provides an effective IIoT‐driven solution for reliable charging demand generation under extreme winter conditions, enhancing energy scheduling and infrastructure resilience for smart city transportation networks.
Zitieren
APA
Lu, X., Li, P., Wang, D., Zhou, X., Liu, M., & Zhao, J. (2026). CEEMDAN ‐ AGNN ‐ GTCN : A Denoising‐Enhanced Spatiotemporal Graph Neural Network for IIoT ‐Based EV Charging Demand Forecasting on Icy and Snowy Roads. Internet Technology Letters, 9(4). https://doi.org/10.1002/itl2.70267
Vancouver
Lu X, Li P, Wang D, Zhou X, Liu M, Zhao J. CEEMDAN ‐ AGNN ‐ GTCN : A Denoising‐Enhanced Spatiotemporal Graph Neural Network for IIoT ‐Based EV Charging Demand Forecasting on Icy and Snowy Roads. Internet Technology Letters. 2026;9(4). doi:10.1002/itl2.70267
BibTeX
@article{Lu2026,
author = {Lu, Xiaoming and Li, Pengfei and Wang, Dingheng and Zhou, Xuelian and Liu, Meihan and Zhao, Jingjie},
title = {CEEMDAN ‐ AGNN ‐ GTCN : A Denoising‐Enhanced Spatiotemporal Graph Neural Network for IIoT ‐Based EV Charging Demand Forecasting on Icy and Snowy Roads},
journal = {Internet Technology Letters},
year = {2026},
volume = {9},
number = {4},
doi = {10.1002/itl2.70267}
}
EUVIMED