2026· IEEE Open Journal of Intelligent Transportation Systems· Vol 7, pp. 2006-2021· 0 citations· 45 references
TL;DR
A novel Adaptive Parameter Coordination Dynamic Hypergraph Spatio-Temporal Prediction (ADH-STPC) framework for decentralized learning, which provides the core topology for adaptive collaboration, while DHCGA captures spatio-temporal traffic features, and together they drive the model to achieve optimal performance.
Abstract
Traffic flow prediction is a critical task in Intelligent Transportation Systems (ITS), essential for alleviating traffic congestion and optimizing resource allocation. While graph neural network-based methods have achieved significant success in modeling pairwise relationships between nodes, they often struggle to capture the dynamic, high-order spatio-temporal dependencies inherent in traffic networks. Furthermore, in distributed real-world scenarios, data heterogeneity among different clients often leads to negative transfer in traditional federated learning methods, thereby limiting model performance. To address these challenges, this paper proposes a novel Adaptive Parameter Coordination Dynamic Hypergraph Spatio-Temporal Prediction (ADH-STPC) framework for decentralized learning. This framework centers on the Client-Adaptive Parameter Sharing (CAPS) module, which coordinates decentralized collaboration by quantifying the similarity of global and local feature distributions, ensuring that parameter propagation occurs only between compatible clients, thereby suppressing negative propagation. Additionally, to accurately predict changes in traffic flow, we designed the Dynamic Hypergraph Convolution Architecture (DHCGA), which captures higher-order nonlinear dependencies by adaptively evolving hypergraph structures, thereby uncovering latent relationships among features. CAPS provides the core topology for adaptive collaboration, while DHCGA captures spatio-temporal traffic features, and together they drive the model to achieve optimal performance. Extensive experiments on two large-scale real-world traffic datasets validate the effectiveness of this framework. Compared to existing state-of-the-art baseline models, ADH-STPC demonstrates superior performance: achieving significant reductions across all key metrics on the BJSubway dataset, including a 15.1% decrease in MAE, a 15.2% reduction in RMSE, and a 7.1% improvement in MAPE.
Experiments on controlled synthetic data and the Q-Traffic real-world dataset demonstrate that the proposed framework improves predictive accuracy, cross-client stability, and robustness under heterogeneous federated traffic scenarios.
Zhi-Cheng Wang, Tao Zhang, Yi-Meng Zhu et al.· Applied Sciences· 0 citations
This framework introduces an adaptive graph learning module that dynamically infers meaningful connectivity relationships among traffic sensors—not relying on fixed geographic or distance-based assumptions—but instead leveraging real-time traffic correlations and node-level embeddings, enabling effective modeling of both localized spatial interactions and multi-scale temporal dependencies across varying prediction horizons.
Zhengxu Luan, Huan Wang, Miaobowen Wang et al.· Computers and artificial int...· 0 citations
An improved Transformer prediction model that integrates a graph convolutional network (GCN) and a self-attention mechanism is proposed for traffic flow prediction, combining temporal self-attention and learnable temporal encoding to capture both long-term traffic evolution patterns and sudden fluctuations.
Jin Zhang, Feng-Min Tan, Wei Bai et al.· Italian National Conference...· 0 citations
Tourist flow modeling and prediction in scenic areas is highly complex, exhibiting significant spatiotemporal dependence and being influenced by various external factors. Traditional models struggle to simultaneously characterize complex spatiotemporal relationships and integrate diverse external information, posing challenges for intelligent resource management and dynamic information scheduling in large-scale networked systems. To address this, this paper proposes a closed-loop integration framework that combines a particle swarm optimization (PSO)-based Temporal Fusion Transformer-Graph Attention Network (TFT-GAT) prediction model with Deep Double-Q Network (D3QN) scheduling optimization. The framework integrates heterogeneous data such as historical traffic, weather, and social media, achieving adaptive time-varying spatial embedding through a graph attention network (GAT) while capturing long- and short-term dependencies using a Temporal Fusion Transformer (TFT) for both point and quantile prediction. The PSO algorithm performs global optimization of the TFT-GAT hyperparameters, and the resulting prediction outputs together with uncertainty estimates are incorporated into the D3QN to realize closed-loop online capacity allocation based on reinforcement learning. Such a data-driven spatiotemporal modeling strategy also provides methodological insights for dynamic information fusion and adaptive resource scheduling in intelligent electromagnetic sensing and communication environments. Experiments conducted at the Forbidden City in Beijing demonstrate high prediction accuracy with a mean absolute error of 1.5–2.6 people/hour, an average quantile coverage exceeding 82%, and a response time of 140.6 ms. The incorporation of exogenous factors, particularly holidays, improves prediction performance by 18.4%, validating the proposed framework’s robustness in spatiotemporal coupling, uncertainty representation, and real-time scheduling.
X. N. Xie, W. L. Zhang· Advanced Electromagnetics· 0 citations
Accurate traffic forecasting is essential for proactive resource management in edge computing, where service demand evolves dynamically across both space and time. In practical cellular edge systems, traffic exhibits strong spatial correlations among neighboring service regions and long-range temporal dependencies driven by user mobility and application behavior. Existing recurrent forecasting approaches can capture short-term dynamics but often struggle to model long-horizon traffic evolution under non-stationary conditions. To address this challenge, we propose a spatiotemporal graph Transformer framework that jointly models spatial interactions and temporal dependencies for traffic forecasting in edge computing. The framework employs graph neural networks to capture spatial correlations among service regions and leverages Transformer-based self-attention to learn long-range temporal patterns from historical traffic observations. By decoupling spatial representation learning from temporal reasoning, the proposed approach provides an effective mechanism for large-scale spatiotemporal traffic modeling. Extensive experiments on a real-world cellular network dataset demonstrate that the proposed graph Transformer consistently outperforms recurrent graph-based baselines, including GCN-RNN, GCN-LSTM, and GCN-GRU models, across multiple forecasting horizons. The resulting forecasts enable more effective proactive resource provisioning and reduce overload risk compared with reactive management strategies. These results highlight the potential of graph-enhanced attention mechanisms for building intelligent and adaptive edge computing systems.
Traffic forecasting plays a crucial role in real-world applications such as traffic management and urban planning. Recent studies have mainly focused on spatio-temporal graph neural networks (STGNNs) and attention-based methods, which have shown promising results. Nevertheless, both approaches model spatial information implicitly, which limits their ability to generalize across different traffic networks. In this paper, we propose Spatio-Temporal Unified Network (STUNet), a framework to explicitly encode spatial features into unified representations and integrate them with temporal information effectively. To obtain spatial representations explicitly, we design a spatial tokenizer that segments the adjacency matrix of the relation graph into patches to serve as spatial tokens. Furthermore, to effectively integrate spatial and temporal representations, we introduce query-aggregate attention, which simulates the process of tracing upstream and downstream nodes and aggregating their information, thereby capturing complex spatio-temporal dependencies. Extensive experiments on traffic benchmarks demonstrate that STUNet achieves generalization across different traffic networks with competitive performance. Code is available at https://github.com/JimmyChen6/STUNet.
Yujun Chen, Shihao Tu, Wenyu Ding et al.· Proceedings of the 32nd ACM...· 0 citations