EUVIMEDEUVIMED entdecken
Crossref

Transformer-Based Generative Adversarial Network for Traffic Forecasting

Bingyi Liu, Lingtian Yuan, Xun Shao, Enshu Wang, Zhenchang Xia, Weizhen Han, Celimuge Wu

IEEE Transactions on Consumer Electronics · 2025-11 · 71 (4) · 12288-12299

DOI: 10.1109/tce.2025.3601144 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Consumer Electronics, 2025-11. Verfasst von Bingyi Liu und 6 weiteren Autor:innen. Die Arbeit wurde bisher 8-mal zitiert (Crossref) und nennt 45 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Wuhan Science and Technology Joint Project for Building a Strong Transportation Country, Wuhan Science and Technology Project for Key Research and Development.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurate traffic flow forecasting is crucial for enhancing the performance and capabilities of navigation applications. The traffic flow data collected from sensors across widely distributed areas exhibits complex and diverse spatial-temporal correlations, posing a significant challenge for effective capture. Additionally, existing models primarily focus on improving the model architecture to enhance prediction accuracy, while overlooking the importance of training methods, resulting in suboptimal performance. To deal with these problems, we propose a Transformer-based Generative Adversarial Network (TGAN) consisting of a meticulously crafted transformer encoder as the generator and Gated Recurrent Units (GRUs) as the discriminator for adversarial training. Specifically, on the generator side, we propose a Time-Dependent Network (TDN) that integrates 1D-CNNs with varying kernel sizes and multi-head attention mechanisms to effectively capture diverse temporal patterns. For spatial modeling, we introduce a Spatial-Dependent Graph Convolutional Network (SDGCN) designed to model dynamic spatial correlations through a learnable masked adjacency matrix. Moreover, criticality embedding is introduced to emphasize the crucial nodes within the road network, while distance embedding is designed to enhance the model’s spatial awareness. On the discriminator side, we concatenate the predicted data from the generator with historical data as input and modify the training loss, thereby ensuring the consistency of the predictions. Subsequently, the generator further improves prediction accuracy through adversarial training with the discriminator. We conduct extensive experiments on four real-world traffic datasets. The experimental results show that the proposed TGAN outperforms the baselines in prediction accuracy.

Förderung

  • National Natural Science Foundation of China
  • Wuhan Science and Technology Joint Project for Building a Strong Transportation Country
  • Wuhan Science and Technology Project for Key Research and Development
  • State Key Laboratory of Integrated Services Networks
  • Natural Science Foundation of Hubei Province
  • JSPS KAKENHI

Zitieren

APA

Liu, B., Yuan, L., Shao, X., Wang, E., Xia, Z., Han, W., & Wu, C. (2025). Transformer-Based Generative Adversarial Network for Traffic Forecasting. IEEE Transactions on Consumer Electronics, 71(4), 12288-12299. https://doi.org/10.1109/tce.2025.3601144

Vancouver

Liu B, Yuan L, Shao X, Wang E, Xia Z, Han W, et al. Transformer-Based Generative Adversarial Network for Traffic Forecasting. IEEE Transactions on Consumer Electronics. 2025;71(4):12288-12299. doi:10.1109/tce.2025.3601144

BibTeX
@article{Liu2025,
  author = {Liu, Bingyi and Yuan, Lingtian and Shao, Xun and Wang, Enshu and Xia, Zhenchang and Han, Weizhen and Wu, Celimuge},
  title = {Transformer-Based Generative Adversarial Network for Traffic Forecasting},
  journal = {IEEE Transactions on Consumer Electronics},
  year = {2025},
  volume = {71},
  number = {4},
  pages = {12288--12299},
  doi = {10.1109/tce.2025.3601144}
}
Kostenlos starten →