交通流变量预测问题的研究现状

Forecasting the traffic flows is a critical issue for researchers and practitioners in the field of transportation. However, it is very challenging since the traffic flows usually show high nonlinearities and complex patterns. Most existing traffic flow prediction methods, lacking abilities of modeling the dynamic spatial-temporal correlations of traffic data, thus cannot yield satisfactory prediction results. In this paper, we propose a novel attention based spatial-temporal graph convolutional network (ASTGCN) model to solve traffic flow forecasting problem. ASTGCN mainly consists of three independent components to respectively model three temporal properties of traffic flows, i.e., recent, daily-periodic and weekly-periodic dependencies. More specifically, each component contains two major parts: 1) the spatial-temporal attention mechanism to effectively capture the dynamic spatialtemporal correlations in traffic data; 2) the spatial-temporal convolution which simultaneously employs graph convolutions to capture the spatial patterns and common standard convolutions to describe the temporal features. The output of the three components are weighted fused to generate the final prediction results. Experiments on two real-world datasets from the Caltrans Performance Measurement System (PeMS) demonstrate that the proposed ASTGCN model outperforms the state-of-the-art baselines.

Traffic prediction is critical for optimizing travel scheduling and enhancing public safety, yet the complex spatial and temporal dynamics within traffic data present significant challenges for accurate forecasting. In this paper, we introduce a novel model, the Spatiotemporal-aware Trend-Seasonality Decomposition Network (STDN). This model begins by constructing a dynamic graph structure to represent traffic flow and incorporates novel spatio-temporal embeddings to jointly capture global traffic dynamics. The representations learned are further refined by a specially designed trend-seasonality decomposition module, which disentangles the trend-cyclical component and seasonal component for each traffic node at different times within the graph. These components are subsequently processed through an encoder-decoder network to generate the final predictions. Extensive experiments conducted on real-world traffic datasets demonstrate that STDN achieves superior performance with remarkable computation cost. Furthermore, we have released a new traffic dataset named JiNan, which features unique inner-city dynamics, thereby enriching the scenario comprehensiveness in traffic prediction evaluation.

Spatial-temporal data forecasting of traffic flow is a challenging task because of complicated spatial dependencies and dynamical trends of temporal pattern between different roads. Existing frameworks usually utilize given spatial adjacency graph and sophisticated mechanisms for modeling spatial and temporal correlations. However, limited representations of given spatial graph structure with incomplete adjacent connections may restrict effective spatial-temporal dependencies learning of those models. Furthermore, existing methods were out at elbows when solving complicated spatial-temporal data: they usually utilize separate modules for spatial and temporal correlations, or they only use independent components capturing localized or global heterogeneous dependencies. To overcome those limitations, our paper proposes a novel Spatial-Temporal Fusion Graph Neural Networks (STFGNN) for traffic flow forecasting. First, a data-driven method of generating “temporal graph” is proposed to compensate several genuine correlations that spatial graph may not reflect. STFGNN could effectively learn hidden spatial-temporal dependencies by a novel fusion operation of various spatial and temporal graphs, treated for different time periods in parallel. Meanwhile, by integrating this fusion graph module and a novel gated convolution module into a unified layer parallelly, STFGNN could handle long sequences by learning more spatial-temporal dependencies with layers stacked. Experimental results on several public traffic datasets demonstrate that our method achieves state-of-the-art performance consistently than other baselines.

Spatial-temporal forecasting has attracted tremendous attention in a wide range of applications, and traffic flow prediction is a canonical and typical example. The complex and long-range spatial-temporal correlations of traffic flow bring it to a most intractable challenge. Existing works typically utilize shallow graph convolution networks (GNNs) and temporal extracting modules to model spatial and temporal dependencies respectively. However, the representation ability of such models is limited due to: (1) shallow GNNs are incapable to capture long-range spatial correlations, (2) only spatial connections are considered and a mass of semantic connections are ignored, which are of great importance for a comprehensive understanding of traffic networks. To this end, we propose Spatial-Temporal Graph Ordinary Differential Equation Networks (STGODE).1 Specifically, we capture spatial-temporal dynamics through a tensor-based ordinary differential equation (ODE), as a result, deeper networks can be constructed and spatial-temporal features are utilized synchronously. To understand the network more comprehensively, semantical adjacency matrix is considered in our model, and a well-design temporal dialated convolution structure is used to capture long term temporal dependencies. We evaluate our model on multiple real-world traffic datasets and superior performance is achieved over state-of-the-art baselines.

Traffic flow forecasting is a classical spatio-temporal data mining problem with many real-world applications. Recently, various methods based on Graph Neural Networks (GNN) have been proposed for the problem and achieved impressive prediction performance. However, we argue that the majority of existing methods disregarding the importance of certain nodes (referred to as pivotal nodes) that naturally exhibit extensive connections with multiple other nodes. Predicting on pivotal nodes poses a challenge due to their complex spatio-temporal dependencies compared to other nodes. In this paper, we propose a novel GNN-based method called Spatio-Temporal Pivotal Graph Neural Networks (STPGNN) to address the above limitation. We introduce a pivotal node identification module for identifying pivotal nodes. We propose a novel pivotal graph convolution module, enabling precise capture of spatio-temporal dependencies centered around pivotal nodes. Moreover, we propose a parallel framework capable of extracting spatio-temporal traffic features on both pivotal and non-pivotal nodes. Experiments on seven real-world traffic datasets verify our proposed method's effectiveness and efficiency compared to state-of-the-art baselines.

Accurate traffic forecasting is a fundamental problem in intelligent transportation systems and learning long-range traffic representations with key information through spatiotemporal graph neural networks (STGNNs) is a basic assumption of current traffic flow prediction models. However, due to structural limitations, existing STGNNs can only utilize short-range traffic flow data; therefore, the models cannot adequately learn the complex trends and periodic features in traffic flow. Besides, it is challenging to extract the key temporal information from the long historical traffic series and obtain a compact representation. To solve the above problems, we propose a novel LSTTN (Long-Short Term Transformer-based Network) framework comprehensively considering the long- and short-term features in historical traffic flow. First, we employ a masked subseries Transformer to infer the content of masked subseries from a small portion of unmasked subseries and their temporal context in a pretraining manner, forcing the model to efficiently learn compressed and contextual subseries temporal representations from long historical series. Then, based on the learned representations, long-term trend is extracted by using stacked 1D dilated convolution layers, and periodic features are extracted by dynamic graph convolution layers. For the difficulties in making time-step level prediction, LSTTN adopts a short-term trend extractor to learn fine-grained short-term temporal features. Finally, LSTTN fuses the long-term trend, periodic features and short-term features to obtain the prediction results. Experiments on four real-world datasets show that in 60-minute-ahead long-term forecasting, the LSTTN model achieves a minimum improvement of 5.63\% and a maximum improvement of 16.78\% over baseline models. The source code is available at https://github.com/GeoX-Lab/LSTTN.

No abstract available

Traffic flow Forecasting is essential in intelligent transportation systems. Although graph neural networks perform well with non-Euclidean traffic data, they exhibit limits in accurately capturing complex spatio-temporal dependencies. Most current methodologies assume instantaneous propagation of traffic flow information, neglecting the delay in information propagation between nodes. Moreover, traditional graph convolutional models capture spatial dependencies using static adjacency matrices, overlooking the dynamic correlations among nodes over time. To address these challenges, this paper proposes a delay propagation spatio-temporal graph convolutional network (DPSTGC) for traffic prediction. By incorporating the delay propagation of traffic flow information into a directed graph of the road network, this model accurately captures the spatio-temporal dependencies of traffic flows. To refine the relationships between nodes, adaptive graph convolution network is employed to learn the dynamic correlation between nodes. The alternating application of temporal gated convolutions and spatial structure significantly enhances the model’s capacity to interpret spatio-temporal information. Furthermore, the effectiveness of the Delay-aware Directed Graph Attention (DDGA) is further interpreted from a causality perspective. Finally, the proposed model is evaluated on four real-world traffic datasets for experimental validation. Experimental results demonstrate that DPSTGC proficiently captures spatio-temporal information and achieves excellent performance.

Traffic flow forecasting task plays an essential role in intelligent transportation systems. Accurately capturing the intricate spatio-temporal dependencies in traffic network signals is the core of precise prediction. Recently, a paradigm that models spatio-temporal dependencies through graph neural networks and time series models has become one of the most promising methods to solve this problem. However, existing methods still have limitations due to ineffectively modeling dynamic spatial dependencies and high time and space complexity. To address these issues, we propose a simplifying and powerful general spatio-temporal traffic flow forecasting model called LightST. Specifically, LightST first embeds temporal covariates and spatial position information to enhance the spatio-temporal modeling capabilities. Then, stacked temporal linear layers are introduced to capture temporal dependencies efficiently. Finally,we propose a concise adaptive spatio-temporal embedding graph convolution method to extract implicit spatial dependencies over time via dynamic graph convolution with adaptive spatio-temporal embedding graph generation. Extensive experiment results on four public traffic flow datasets demonstrate the superiority of our LightST concerning computational efficiency and prediction performance.

No abstract available

No abstract available

ABSTRACT Accurate real-time traffic flow forecasting has been a challenge due to the complex spatial–temporal dependencies and uncertainties associated with the dynamic changes in traffic flow. To overcome this problem, a traffic flow forecasting model based on an Augmented Multi-Component Recurrent Graph Attention Network (AMR-GAT) is proposed in this paper to model the spatial–temporal correlations and periodic offset of traffic flows. This paper introduces an augmented multi-component module to address periodic temporal offset in traffic flow forecasting. It proposes an encoder-decoder architecture combining 1D convolution and LSTM via a Temporal Correlation Learner (TCL) to capture temporal characteristics, while a Graph Attention Network (GAT) handles spatial features. TCL and GAT are integrated to manage spatial-temporal correlations, and the decoder uses TCL and convolutional neural networks to generate high-dimensional representations based on spatial-temporal sequences. Experiments on two datasets demonstrate superior prediction performance of the proposed AMR-GAT model.

No abstract available

No abstract available

Traffic forecasting is a fundamental task in transportation research, however, the scope of current research has mainly focused on a single data modality of loop detectors. Recently, the advances in Artificial Intelligence and drone technologies have enabled novel solutions to efficient, accurate, and flexible aerial observations of urban traffic. As a promising traffic monitoring approach, drones can create an accurate multi-sensor mobility observatory for large-scale urban networks, when combined with existing infrastructure. Therefore, this paper investigates urban traffic prediction from a novel perspective, where multiple input sources, i.e., loop detectors and drones, are utilized to predict both the segment-level traffic of all road segments and the regional traffic. A simple yet effective graph-based model, Hierarchical Multi-Source Neural Network (HiMSNet), is proposed to integrate multiple data modalities and learn spatio-temporal correlations. Detailed analysis shows that predicting accurate segment-level speed is more challenging than the regional speed, especially under high-demand scenarios with heavier congestion and varying traffic dynamics. Utilizing both drone and loop detector data, the prediction accuracy can be improved compared to single-modality cases, when the sensors have lower coverage and are subject to noise. Our simulation study based on vehicle trajectories in a real urban road network has highlighted the value of integrating drones in traffic forecasting and monitoring.

Traffic flow is the most direct indicator of traffic conditions, and accurate prediction of traffic flow is a key challenge for scholars in the field of intelligent transportation. However, traffic flow displays significant nonlinearity, dynamic changes, spatiotemporal dependencies, and most existing methods overlook the influence of road topology on the spatiotemporal properties of traffic flow, creating substantial challenges for traffic flow prediction. This paper proposes a graph convolutional traffic flow prediction model based on adaptive spatiotemporal attention. Initially, the model adaptively adjusts spatiotemporal weight distribution using a meticulously designed spatiotemporal attention mechanism, effectively capturing dynamic spatiotemporal correlations in traffic data. Subsequently, it integrates graph convolutional neural networks (GCNs) with long short-term memory (LSTM) networks to capture the spatiotemporal characteristics of traffic data. Additionally, a GCN is designed to capture the spatial topological relationships of the road network. Finally, a novel fusion mechanism is introduced to integrate the spatiotemporal features of traffic data with the spatial topological relationships of roads, aiming to achieve accurate predictions. Experimental results demonstrate that the model proposed in this paper outperforms six selected baseline methods.

Most current methods use spatial–temporal graph neural networks (STGNNs) to analyze complex spatial–temporal information from traffic data collected from hundreds of sensors. STGNNs combine graph neural networks (GNNs) and sequence models to create hybrid structures that allow for the two networks to collaborate. However, this collaboration has made the model increasingly complex. This study proposes a framework that relies solely on original Transformer architecture and carefully designs embeddings to efficiently extract spatial–temporal dependencies in traffic flow. Additionally, we used pre-trained language models to enhance forecasting performance. We compared our new framework with current state-of-the-art STGNNs and Transformer-based models using four real-world traffic datasets: PEMS04, PEMS08, METR-LA, and PEMS-BAY. The experimental results demonstrate that our framework outperforms the other models in most metrics.

With the significant increase in the number of motor vehicles, road-related issues, such as traffic congestion and accidents, have also escalated. The development of an accurate and efficient traffic flow forecasting model is essential for helping car owners plan their journeys. Despite advancements in forecasting models, there are three remaining issues: 1) failing to effectively use cyclical data; 2) failing to adequately capture spatial dependencies; and 3) high-time complexity and memory usage. To tackle the aforementioned challenges, we present a novel spatial–temporal graph attention gated recurrent transformer network (STGAGRTN) for traffic flow forecasting. Specifically, the use of a spatial transformer module allows for the extraction of dynamic spatial dependencies among individual nodes, going beyond the limitation of only considering neighboring nodes. Subsequently, we propose a temporal transformer to extract periodic information from traffic data and capture long-term dependencies. Additionally, we utilize two additional classical techniques to complement the aforementioned modules for extracting characteristics. By incorporating comprehensive spatial-temporal characteristics into our model, we can accurately predict multiple nodes simultaneously. Finally, we have successfully optimized the computational complexity of the transformer module from $\mathcal {O}(n^{2})$ to $\mathcal {O}(n \log n)$ . Our model has undergone extensive testing on four authentic data sets, providing compelling evidence of its superior predictive capabilities.

Traffic flow forecasting is essential and challenging to intelligent city management and public safety. In this paper, we attempt to use a pure self-attention method in traffic flow forecasting. However, when dealing with input sequences, including large-scale regions’ historical records, it is difficult for the self-attention mechanism to focus on the most relevant ones for forecasting. To address this issue, we design a progressive space-time self-attention mechanism named ProSTformer, which can reduce self-attention computation times from thousands to tens. Our design is based on two pieces of prior knowledge in the traffic flow forecasting literature: (i) spatiotemporal dependencies can be factorized into spatial and temporal dependencies; (ii) adjacent regions have more influences than distant regions, and temporal characteristics of closeness, period and trend are more important than crossed relations between them. Our ProSTformer has two characteristics. First, each block in ProSTformer highlights the unique dependencies, ProSTformer progressively focuses on spatial dependencies from local to global regions, on temporal dependencies from closeness, period and trend to crossed relations between them, and on external dependencies such as weather conditions, temperature and day-of-week. Second, we use the Tensor Rearranging technique to force the model to compute self-attention only to adjacent regions and to the unique temporal characteristic. Then, we use the Patch Merging technique to greatly reduce self-attention computation times to distant regions and crossed temporal relations. We evaluate ProSTformer on two traffic datasets and find that it performs better than sixteen baseline models. The code is available at https://github.com/yanxiao1930/ProSTformer_code/tree/main.

Graph convolutional networks (GCN) have been applied in the traffic flow forecasting tasks with the graph capability in describing the irregular topology structures of road networks. However, GCN based traffic flow forecasting methods often fail to simultaneously capture the short-term and long-term temporal relations carried by the traffic flow data, and also suffer the over-smoothing problem. To overcome the problems, we propose a hierarchical traffic flow forecasting network by merging newly designed the long-term temporal Transformer network (LTT) and the spatio-temporal graph convolutional networks (STGC). Specifically, LTT aims to learn the long-term temporal relations among the traffic flow data, while the STGC module aims to capture the short-term temporal relations and spatial relations among the traffic flow data, respectively, via cascading between the one-dimensional convolution and the graph convolution. In addition, an attention fusion mechanism is proposed to combine the long-term with the short-term temporal relations as the input of the graph convolution layer in STGC, in order to mitigate the over-smoothing problem of GCN. Experimental results on three public traffic flow datasets prove the effectiveness and robustness of the proposed method.

No abstract available

Accurate traffic flow forecasting is a crucial component of intelligent transportation systems, playing a pivotal role in enhancing transportation intelligence. The integration of Graph Neural Networks (GNNs) and Transformers in traffic flow forecasting has gained significant adoption for enhancing prediction accuracy. Yet, the complex spatial and temporal dependencies present in traffic data continue to pose substantial challenges: (1) Most GNN-based methods assume that the graph structure reflects the actual dependencies between nodes, overlooking the complex dependencies present in the real-world context. (2) Standard time-series models are unable to effectively model complex temporal dependencies, hindering prediction accuracy. To tackle these challenges, the authors propose a novel Transformer-based Spatiotemporal Graph Diffusion Convolution Network (TSGDC) for Traffic Flow Forecasting, which leverages graph diffusion and transformer to capture the complexity and dynamics of spatial and temporal patterns, thereby enhancing prediction performance. The authors designed an Efficient Channel Attention (ECA) that learns separately from the feature dimensions collected by traffic sensors and the temporal dimensions of traffic data, aiding in spatiotemporal modeling. Chebyshev Graph Diffusion Convolution (GDC) is used to capture the complex dependencies within the spatial distribution. Sequence decomposition blocks, as internal operations of transformers, are employed to gradually extract long-term stable trends from hidden complex variables. Additionally, by integrating multi-scale dependencies, including recent, daily, and weekly patterns, accurate traffic flow predictions are achieved. Experimental results on various public datasets show that TSGDC outperforms conventional traffic forecasting models, particularly in accuracy and robustness.

Traffic flow prediction is crucial for efficient traffic management. It involves predicting vehicle movement patterns to reduce congestion and enhance traffic flow. However, the highly non-linear and complex patterns commonly observed in traffic flow pose significant challenges for this task. Current Graph Neural Network (GNN) models often construct shallow networks, which limits their ability to extract deeper spatio-temporal representations. Neural ordinary differential equations for traffic prediction address over-smoothing but require significant computational resources, leading to inefficiencies, and sometimes deeper networks may lead to poorer predictions for complex traffic information. In this study, we propose an Adaptive Decision spatio-temporal Neural Ordinary Differential Network, which can adaptively determine the number of layers of ODE according to the complexity of traffic information. It can solve the over-smoothing problem better, improving overall efficiency and prediction accuracy. In addition, traditional temporal convolution methods make it difficult to deal with complex and variable traffic time information with a large time span. Therefore, we introduce a multi-kernel temporal dynamic expansive convolution to handle the traffic time information. Multi-kernel temporal dynamic expansive convolution employs a dynamic dilation strategy, dynamically adjusting the network's receptive field across levels, effectively capturing temporal dependencies, and can better adapt to the changing time data of traffic information. Additionally, multi-kernel temporal dynamic expansive convolution integrates multi-scale convolution kernels, enabling the model to learn features across diverse temporal scales. We evaluated our proposed method on several real-world traffic datasets. Experimental results show that our method outperformed state-of-the-art benchmarks.

Traffic flow forecasting is a crucial task in transportation management and planning. The main challenges for traffic flow forecasting are that (1) as the length of prediction time increases, the accuracy of prediction will decrease; (2) the predicted results greatly rely on the extraction of temporal and spatial dependencies from the road networks. To overcome the challenges mentioned above, we propose a multi-channel spatial-temporal transformer model for traffic flow forecasting, which improves the accuracy of the prediction by fusing results from different channels of traffic data. Our approach leverages graph convolutional network to extract spatial features from each channel while using a transformer-based architecture to capture temporal dependencies across channels. We introduce an adaptive adjacency matrix to overcome limitations in feature extraction from fixed topological structures. Experimental results on six real-world datasets demonstrate that introducing a multi-channel mechanism into the temporal model enhances performance and our proposed model outperforms state-of-the-art models in terms of accuracy.

No abstract available

No abstract available

Traffic flow prediction is the foundation of traffic scheduling and a major component of intelligent transportation systems (ITSs). Accurate traffic flow prediction is crucial for numerous real-world traffic applications. However, the complex and dynamic nature of spatiotemporal correlations in traffic flow data poses obstacles to this task. Another important factor hindering model improvement is the scarcity of spatio-temporal data. Meanwhile, most existing studies assume the traffic data collected by sensors in real scenarios is completely correct and trustworthy, thereby neglecting the anomalies and incorrect data. To address these issues, we propose the contrastive learning-based adaptive graph fusion convolution network with residual-enhanced decomposition strategy (CDAGF) for traffic flow forecasting, which performs a simple yet effective graph augmentation mechanism imposed only on the learned graphs and reserves the explicit graph to mitigate the data scarcity problem, and designs a negative filter to assign negative pairs based on timestamp information. Besides, CDAGF separates anomalous signals from valid traffic signals through a residual-enhanced decomposition strategy to weaken the impact of anomalies and further improve the prediction accuracy. Moreover, CDAGF generates static graph and dynamic graphs based on traffic signals as well as timestamp information and fuses the learned graphs with the explicit graph to model temporal and spatial characteristics dynamically and adaptively. The experimental results on five popular public data sets demonstrate that CDAGF achieves state-of-the-art performance with 18.29 mean absolute error (MAE) on the PEMS04 data set.

Because of the random volatility of traffic data, short-term traffic flow forecasting has always been a problem that needs to be further researched. We developed a short-term traffic flow forecasting approach by applying a secondary decomposition strategy and CNN–Transformer model. Firstly, traffic flow data are decomposed by using a Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) algorithm, and a series of intrinsic mode functions (IMFs) are obtained. Secondly, the IMF1 obtained from the CEEMDAN is further decomposed into some sub-series by using Variational Mode Decomposition (VMD) algorithm. Thirdly, the CNN–Transformer model is established for each IMF separately. The CNN model is employed to extract local spatial features, and then the Transformer model utilizes these features for global modeling and long-term relationship modeling. Finally, we obtain the final results by superimposing the forecasting results of each IMF component. The measured traffic flow dataset of urban expressways was used for experimental verification. The experimental results reveal the following: (1) The forecasting performance achieves remarkable improvement when considering secondary decomposition. Compared with the VMD-CNN–Transformer, the CEEMDAN-VMD-CNN–Transformer method declined by 25.84%, 23.15% and 22.38% in three-step-ahead forecasting in terms of MAPE. (2) It has been proven that our proposed CNN–Transformer model could achieve more outstanding forecasting performance. Compared with the CEEMDAN-VMD-CNN, the CEEMDAN-VMD-CNN–Transformer method declined by 13.58%, 11.88% and 11.10% in three-step-ahead forecasting in terms of MAPE.

Accurate spatial-temporal traffic flow forecasting is crucial in aiding traffic managers in implementing control measures and assisting drivers in selecting optimal travel routes. Traditional deep-learning based methods for traffic flow forecasting typically rely on historical data to train their models, which are then used to make predictions on future data. However, the performance of the trained model usually degrades due to the temporal drift between the historical and future data. To make the model trained on historical data better adapt to future data in a fully online manner, this paper conducts the first study of the online test-time adaptation techniques for spatial-temporal traffic flow forecasting problems. To this end, we propose an Adaptive Double Correction by Series Decomposition (ADCSD) method, which first decomposes the output of the trained model into seasonal and trend-cyclical parts and then corrects them by two separate modules during the testing phase using the latest observed data entry by entry. In the proposed ADCSD method, instead of fine-tuning the whole trained model during the testing phase, a lite network is attached after the trained model, and only the lite network is fine-tuned in the testing process each time a data entry is observed. Moreover, to satisfy that different time series variables may have different levels of temporal drift, two adaptive vectors are adopted to provide different weights for different time series variables. Extensive experiments on four real-world traffic flow forecasting datasets demonstrate the effectiveness of the proposed ADCSD method. The code is available at https://github.com/Pengxin-Guo/ADCSD

Traffic flow forecasting is integral to transportation to avoid traffic accidents and congestion. Due to the heterogeneous and nonlinear nature of the data, traffic flow prediction is facing challenges. Existing models only utilize plain historical data for prediction. Inadequate use of temporal information has become a key problem in current forecasting. To address the problem, we must effectively analyze the influence of time periods while integrating the distinct characteristics of traffic flow across various time granularities. This paper proposed a multi-granularity temporal embedding Transformer network, namely MGTEFormer. An embedding input adeptly merges complex temporal embeddings, a temporal encoder to consolidate rich temporal information, and a spatial encoder to discern inherent spatial characteristics between different sensors. The combined embeddings are fed into the attention mechanism’s encoder, culminating in prediction results obtained through a linear regression layer. Temporal embedding can help by fussing the period and various temporal granularities into plain historical traffic flow that can be learned adequately, reducing the loss of time information. Experimental analyses and ablation studies conducted on real traffic datasets consistently attest to the superior performance of the MGTEFormer. Our approach reduces the mean absolute error of the original models by less than 1.7%. Extensive experiments demonstrate the superiority of the proposed MGTEFormer over existing benchmarks.

Traffic flow forecasting is crucial for improving urban traffic management and reducing resource consumption. Accurate traffic conditions prediction requires capturing the complex spatial-temporal dependencies inherent in traffic data. Traditional spatial-temporal graph modeling methods often rely on fixed road network structures, failing to account for the dynamic spatial correlations that vary over time. To address this, we propose a Transformer-Enhanced Adaptive Graph Convolutional Network (TEA-GCN) that alternately learns temporal and spatial correlations in traffic data layer-by-layer. Specifically, we design an adaptive graph convolutional module to dynamically capture implicit road dependencies at different time levels and a local-global temporal attention module to simultaneously capture long-term and short-term temporal dependencies. Experimental results on two public traffic datasets demonstrate the effectiveness of the proposed model compared to other state-of-the-art traffic flow prediction methods.

Traffic flow forecasting aims to predict future traffic flows based on the historical traffic conditions and the road network. It is an important problem in intelligent transportation systems, with a plethora of methods been proposed. Existing efforts mainly focus on capturing and utilizing spatio-temporal dependencies to predict future traffic flows. Though promising, they fall short in adapting to test-time environmental changes of traffic conditions. To tackle this challenge, we propose to introduce large language models (LLMs) to help traffic flow forecasting and design a novel method named Large Language Model Enhanced Traffic Flow Predictor (LEAF). LEAF adopts two branches, capturing different spatio-temporal relations using graph and hypergraph structures respectively. The two branches are first pre-trained individually, and during test-time, they yield different predictions. Based on these predictions, a large language model is used to select the most likely result. Then, a ranking loss is applied as the learning objective to enhance the prediction ability of the two branches. Extensive experiments on several datasets demonstrate the effectiveness of the proposed LEAF.

Multiple federated learning (FL) methods are proposed for traffic flow forecasting (TFF) to avoid heavy-transmission and privacy-leaking concerns resulting from the disclosure of raw data in centralized methods. However, these FL methods adopt offline learning which may yield subpar performance, when concept drift occurs, i.e., distributions of historical and future data vary. Online learning can detect concept drift during model training, thus more applicable to TFF. Nevertheless, the existing federated online learning method for TFF fails to efficiently solve the concept drift problem and causes tremendous computing and communication overhead. Therefore, we propose a novel method named Resource-Efficient Federated Online Learning (REFOL) for TFF, which guarantees prediction performance in a communication-lightweight and computation-efficient way. Specifically, we design a data-driven client participation mechanism to detect the occurrence of concept drift and determine clients’ participation necessity. Subsequently, we propose an adaptive online optimization strategy, which guarantees prediction performance and meanwhile avoids meaningless model updates. Then, a graph convolution-based model aggregation mechanism is designed, aiming to assess participants’ contribution based on spatial correlation without importing extra communication and computing consumption on clients. Finally, we conduct extensive experiments on real-world datasets to demonstrate the superiority of REFOL in terms of prediction improvement and resource economization.

Traffic flow forecasting is a crucial yet complex task due to the intricate spatial-temporal correlations arising from road interactions. Recent methods model these interactions using message-passing Graph Convolution Networks (GCNs), which work for homophily graphs where connected nodes primarily exhibit close observations. However, relying solely on homophily graphs presents inherent limitations in traffic modeling, as road interactions can yield not only close but also distant observations over time, revealing diverse and dynamic node-wise correlations. We designate this phenomenon as homophily-heterophily dynamics, which has been largely overlooked in previous works. To address this gap, we propose a homophily-heterophily Spatial-Temporal Graph Convolution Network (H2STGCN) that exploits both homophily and heterophily components in the spatial-temporal domain. Specifically, we first adopt time-related node attributes to disentangle the diverse and dynamic node-wise relations across time, thereby obtaining homophily and heterophily Spatial-Temporal Graphs (STGs), which provide comprehensive insights into road interactions. Subsequently, we construct dual information propagation branches, each outfitted with a specific type of STG, to exploit multiple ranges of spatial-temporal correlations from distinct perspectives through dilated causal spatial-temporal graph convolution operations on STGs. Additionally, we introduce a Graph Collaborative Learning Module (GCLM) to capture the complementary information of these two branches via mutual information transfer. Experimental evaluation on four real-world traffic datasets reveals that our model outperforms state-of-the-art methods.

Graph neural networks integrating contrastive learning have attracted growing attention in urban traffic flow forecasting. However, most existing graph contrastive learning methods do not perform well in capturing local–global spatial dependencies or designing contrastive learning schemes for both spatial and temporal dimensions. We argue that these methods can not well extract the spatial-temporal features and are easily affected by data noise. In light of these challenges, this paper proposes an innovative Urban Spatial-Temporal Graph Contrastive Learning framework (UrbanGCL) to improve the accuracy of urban traffic flow forecasting. Specifically, UrbanGCL proposes multi-level data augmentation to address data noise and incompleteness, learn both local and global topology features. The augmented traffic feature matrices and adjacency matrices are then fed into a simple yet effective dual-branch network with shared parameters to capture spatial-temporal correlations within traffic sequences. Moreover, we introduce spatial and temporal contrastive learning auxiliary tasks to alleviate the sparsity of supervision signal and extract the most critical spatial-temporal information. Extensive experimental results on four real-world urban datasets demonstrate that UrbanGCL significantly outperforms other state-of-the-art methods, with the maximum improvement reaching nearly 8.80%.

An essential component of autonomous transportation system management and decision-making is precise and real-time traffic flow forecast. Predicting future traffic conditionsis a difficult undertaking because of the intricate spatio-temporal relationships involved. Existing techniques often employ separate modules to model spatio-temporal features independently, thereby neglecting the temporally and spatially heterogeneous features among nodes. Simultaneously, many existing methods overlook the long-term relationships included in traffic data, subsequently impacting prediction accuracy. We introduce a novel method to traffic flow forecasting based on the combination of the feature-augmented down-sampling dynamic graph convolutional network and multi-head attention mechanism. Our method presents a feature augmentation mechanism to integrate traffic data features at different scales. The subsampled convolutional network enhances information interaction in spatio-temporal data, and the dynamic graph convolutional network utilizes the generated graph structure to better simulate the dynamic relationships between nodes, enhancing the model’s capacity for capturing spatial heterogeneity. Through the feature-enhanced subsampled dynamic graph convolutional network, the model can simultaneously capture spatio-temporal dependencies, and coupled with the process of multi-head temporal attention, it achieves long-term traffic flow forecasting. The findings demonstrate that the ADDGCN model demonstrates superior prediction capabilities on two real datasets (PEMS04 and PEMS08). Notably, for the PEMS04 dataset, compared to the best baseline, the performance of ADDGCN is improved by 2.46% in MAE and 2.90% in RMSE; for the PEMS08 dataset, compared to the best baseline, the ADDGCN performance is improved by 1.50% in RMSE, 3.46% in MAE, and 0.21% in MAPE, indicating our method’s superior performance.

Accurate traffic flow forecasting is important for intelligent traffic management and control. To address the inability of existing methods to simultaneously capture the spatiotemporal dependence of traffic flows and the significant trend differences between predicted and real values, a traffic flow forecasting method based on a Spatiotemporal Interactive Dynamic Adaptive Convolutional Network (STIDAAG) is proposed. First, an interactive learning structure is designed to dynamically aggregate the spatiotemporal characteristics of the hidden nodes in the traffic network. Second, a dynamic adaptive graph generation network is designed based on the current and historical state to further capture the dynamic spatiotemporal characteristics. Finally, the adversarial graph convolutional network is used to optimize the loss for adversarial training to reduce the trend difference between the predicted and true values. The results of the experiment on four publicly available datasets indicate that STIDAAG outperforms both typical and advanced methods in terms of predictive performance.

No abstract available

Over the past decade, major cities have faced significant traffic congestion, accidents, and pollution due to increased vehicle usage, urbanization, and migration. An Intelligent Transportation System (ITS) can enhance transportation planning and alleviate congestion. ITS utilizes traffic prediction models to help prevent traffic bottlenecks, improve mobility and safety, and reduce environmental impacts. However, developing these models involves several challenges, including understanding spatiotemporal nonlinearities, making accurate predictions, minimizing prediction time, and reducing model complexity. Many existing approaches integrate Convolutional Neural Networks (CNNs) and variants of Recurrent Neural Networks (RNNs) to analyze spatially correlated traffic data over time. Nevertheless, these hybrid models often require significant storage space, contain numerous learnable parameters, and involve extensive training, validation, and testing times. To address these challenges, we propose a novel methodology that combines a genetic algorithm (GA) with Random Forest Cross-Validation (RF-CV) to evaluate input features and select the most relevant subset. Additionally, we developed a Multi-Objective Genetic Algorithm (MOGA)-enhanced RNN model to optimize hyperparameters and achieve accurate traffic speed predictions. Our proposed methodology balances the trade-offs between prediction accuracy, model size, and computational efficiency by identifying an optimal set of relevant features and hyperparameters. We evaluated our model using the Performance Measurement System (PeMS)-10 dataset and compared its performance against baseline and advanced models from existing literature. Our model achieved a Mean Absolute Error (MAE) of 0.028993, an $R^{2}$ score of 0.999490, and training, validation, and testing times of 81.64 seconds, 0.15 seconds, and 0.18 seconds, respectively. Additionally, the model size was 203,118 bytes, with 14,617 parameters. A comprehensive comparative study demonstrates that our approach outperforms state-of-the-art models in both prediction accuracy and computational efficiency.

Short-term traffic speed prediction is fundamental to intelligent transportation systems (ITS), and the accuracy of the model largely determines the performance of real-time traffic control and management. In this study, a short-term traffic speed prediction method based on the spatial-temporal analysis of traffic flow and a combined deep-learning model, and a hybrid spatial-temporal feature selection algorithm (STFSA) of a convolutional neural network–gated recurrent unit (CNN-GRU)) is initially developed. Specifically, the STFSA is firstly employed to reconstruct the spatial-temporal matrix of traffic speed based on temporal continuity and spatial characteristics, and then this matrix is considered as the input feature of the prediction model. After this, the nonlinear fitting ability of the CNN is adopted to extract deep features from the convolutional and pooling layers for model training. Finally, by combining the timing and long-range dependence of the captured data with the forward GRU and the reverse GRU, the accuracy of the prediction result is further improved. The validity of the proposed model can be verified by comparing the prediction results with the actual traffic data. Accordingly, in the case study, the performance is compared with various benchmark methods under the same prediction scenario, verifying the superiority of the proposed model.

Urban traffic speed prediction aims to estimate the future traffic speed for improving urban transportation services. Enormous efforts have been made to exploit Graph Neural Networks (GNNs) for modeling spatial correlations and temporal dependencies of traffic speed evolving patterns, regularized by graph topology. While achieving promising results, current traffic speed prediction methods still suffer from ignoring topology-free patterns, which cannot be captured by GNNs. To tackle this challenge, we propose a generic model for enabling the current GNN-based methods to preserve topology-free patterns. Specifically, we first develop a Dual Cross-Scale Transformer (DCST) architecture, including a Spatial Transformer and a Temporal Transformer, to preserve the cross-scale topology-free patterns and associated dynamics, respectively. Then, to further integrate both topology-regularized/-free patterns, we propose a distillation-style learning framework, in which the existing GNN-based methods are considered as the teacher model, and the proposed DCST architecture is considered as the student model. The teacher model would inject the learned topology-regularized patterns into the student model for integrating topology-free patterns. The extensive experimental results demonstrated the effectiveness of our methods.

No abstract available

No abstract available

Accurate and real-time traffic speed prediction remains challenging due to the irregularity and asymmetry of real-traffic road networks. Existing models based on graph convolutional networks commonly use multi-layer graph convolution to extract an undirected static adjacency matrix to map the correlation of nodes, which ignores the dynamic symmetry change of correlation over time and faces the challenge of oversmoothing during training iterations, making it difficult to learn the spatial structure and temporal trend of the traffic network. To overcome the above challenges, we propose a novel multi-head self-attention gated spatiotemporal graph convolutional network (MSGSGCN) for traffic speed prediction. The MSGSGCN model mainly consists of the Node Correlation Estimator (NCE) module, the Time Residual Learner (TRL) module, and the Gated Graph Convolutional Fusion (GGCF) module. Specifically, the NCE module aims to capture the dynamic spatiotemporal correlations between nodes. The TRL module utilizes a residual structure to learn the long-term temporal features of traffic data. The GGCF module relies on adaptive diffusion graph convolution and gated recurrent units to learn the key spatial features of traffic data. Experimental analysis on a pair of real-world datasets indicates that the proposed MSGSGCN model enhances prediction accuracy by more than 4% when contrasted with state-of-the-art models.

No abstract available

Traffic speed prediction is a crucial task for optimizing navigation systems and reducing traffic congestion. Although there have been efforts to improve the accuracy of speed prediction by incorporating auxiliary features, such as traffic flow, weather, and time, types of auxiliary features are limited and their detailed relationships with speed have not been explored yet. In our study, we present the individual spatio-temporal (IST) dependencies on flow and speed, and characterize three types of IST-dependencies with the flow-to-flow, speed-to-speed, and flow-to-speed graphs. Then, we propose Auxiliary feature-aided Attention Network (ARIAN), a novel approach to judiciously learning the degrees of IST-dependencies with the three graphs and predicting the future speed by leveraging various auxiliary features. Through comprehensive experiments using 3 real-world datasets, we validate the superiority of ARIAN over 10 state-of-the-art methods and the effectiveness of each auxiliary feature and each dependency learner in ARIAN.

Advanced technologies, driven by extensive data analysis, support the concept of intelligent cities, which aim to enhance the quality of people’s lives, minimize the consumption of energy, reduce pollution, and promote economic growth. The transportation network is a crucial component of this vision in urbanized cities. However, a massive increase in road traffic poses a significant challenge to achieving this vision. Developing an intelligent transportation system requires accurately predicting the traffic speed. This paper proposes a novel deep stacking-based Ensemble model with a two-layer structure to address the problem of forecasting traffic speed in urbanized transportation networks to solve traffic congestion problems. Firstly, advanced machine learning such as eXtreme Gradient Boosting(XGB), Random Forest(RF), and Extra Tree(ET) as base learners are used to predict short-term traffic speed. In the next phase, the Multi-Layer Perceptron (MLP) as a meta-learner technique, employing various combinations of the aforementioned approaches is used to enhance the accuracy of traffic speed prediction. The proposed stacking-based approach has the capability to analyze, extract, and aggregate various features from primary traffic speed data in order to generate more refined and accurate forecasts. This study used a publicly available dataset of Floating Cars Data collected from real transportation networks for evaluation. Mutual information regression is used as a feature selection technique to obtain the features from the dataset for the training of these models. The performance results are compared with state-of-the-art traffic prediction models. Results show that the proposed stacking-based ensemble strategy outperforms conventional approaches by a large margin such as HA, KNN, SVR, DT, T-GCN, and A3TGCN models. The results demonstrate a notable reduction of 9.71% in RMSE and 15.4% in MAE, indicating enhanced accuracy. Furthermore, our approach achieved a substantial improvement of 13.80% in $\mathbb {R}^{2}$ and 11.64% in EV for the 15-minute prediction horizon.

Predicting traffic speed accurately and in real-time is crucial for the development of smart transportation systems. Given the nonlinear and stochastic nature of vehicle data, integrating diverse spatio-temporal data sources with the Improved Particle Swarm Optimisation (IPSO) offers a promising approach to optimise the Long Short-Term Memory Neural Network (LSTM). Firstly, we enhance the optimisation capabilities of PSO by implementing nonlinear inertial weight and adaptive variation. Secondly, addressing the challenge of selecting the LSTM hyperparameters, the PSO algorithm effectively identifies global optimal solutions for hyperparameter optimisation, ensuring appropriate settings through iterative training. Subsequently, we conduct a case study using multi-source spatio-temporal traffic speed data, comparing our proposed IPSO-LSTM model with traditional neural network prediction models and advanced models. Results from the experiment demonstrate that the IPSO-LSTM model presented in this study addresses issues of parameter selection and inaccurate prediction encountered by traditional LSTM models in traffic state prediction. Moreover, it enhances the model’s ability to capture speed time series dynamics. Notably, in processing complex speed data, our model exhibits superior accuracy and stability in prediction.

No abstract available

Accurate traffic speed prediction is crucial for the guidance and management of urban traffic, which at the same time requires a model with a satisfactory computational burden and memory space in applications. In this paper, we propose a factorized Spatial-Temporal Tensor Graph Convolutional Network for traffic speed prediction. Traffic networks are modeled and unified into a graph tensor that integrates spatial and temporal information simultaneously. We extend graph convolution into tensor space and propose a tensor graph convolution network to extract more discriminating features from spatial-temporal graph data. We further introduce Tucker decomposition and derive a factorized tensor convolution to reduce the computational burden, which performs separate filtering in small-scale space, time, and feature modes. Besides, we can benefit from noise suppression of traffic data when discarding those trivial components in the process of tensor decomposition. Extensive experiments on the three real-world datasets demonstrate that our method is more effective than traditional prediction methods, and achieves state-of-the-art performance.

As the intelligent transportation system has been introduced, traffic speed prediction has become one of the foremost challenging tasks within complex urban road networks. The main idea of this study is to identify links that have a significant impact on the target link and develop a high-performance travel speed prediction model using those links. This study proposes the Extreme gradient boosting model with high importance links (HI-XGB) to predict traffic speed in the urban area. High importance links for predicting the target link speed are selected using Shapley additive explanations. With the selected input features, extreme gradient boosting is used to predict traffic speed. The results show that the performance of the HI-XGB model with one- and 12-time steps ahead achieved 98.5% and 90.7% accuracy, respectively. Feature analysis and link classification analysis are performed to identify the impact of the spatial characteristic on predicted speed. Among the eight features, the speed of the target link at t and the speed change of the target link at t–1 have the most impact on the predicted target link speed. In addition, link classification analysis is performed to identify the impact of the spatial characteristic of the input feature on predicted speed. The result indicates that links other than upstream or downstream could have a greater impact on traffic speed prediction.

No abstract available

No abstract available

Traffic prediction plays a crucial role in an intelligent transportation system (ITS) for enabling advanced transportation management and services. In this paper, we address the problem of multi-step traffic speed prediction, including both short-and long-term predictions. We assert that it is important to consider not just the fixed spatial dependency of the road network (i.e., the connections between road segments) but also the dynamic spatial dependency of traffic within the static topology that intertwines with the temporal evolution of traffic condition across the entire network. We propose a novel deep learning model, named Self-Attention Graph Convolutional Network with Spatial, Sub-spatial and Temporal blocks (SAGCN-SST) model, that specifically capture such complex dynamic spatial-temporal processes. In SAGCN-SST, we integrate self-attention mechanism into graph convolutional networks in a novel framework design while using a sequence-to-sequence model in an encoder-decoder architecture for extracting long-temporal dependency of traffic speed. Two real-world datasets with frequent traffic congestion and accidents from large-scale road networks (i.e., Seattle and Los Angeles) are used to train and test our model. Our experiment results indicate that the proposed deep learning model consistently achieves the most accurate predictions (higher than 98% accuracy on both datasets for the short-and long-term predictions) when compared against well-known existing models in recent literature. The results also indicate that SAGCN-SST is robust against emergent traffic situations.

Speed prediction is a crucial yet complicated task for intelligent transportation systems. The challenge derives from the complex spatiotemporal dependencies of traffic parameters. In the past few years, deep neural networks have achieved the best traffic speed prediction performance. However, most models depend on short-term input sequences to predict short/long-term traffic speed (e.g., predicting speed for the next hour using data from the past hour). These models fail to consider the daily and weekly periodic behavior of traffic. Another problem posed by neural networks is the lack of interpretability as they often operate as “black boxes”. In this paper, an attention-based multi-encoder-decoder (Att-MED) model is proposed to predict traffic speed. The model uses convolutional-LSTMs to capture the spatiotemporal relationship of multiple input sequences, namely short-term, daily and weekly traffic patterns. The model also employs an LSTM to model the output predictions sequentially. Furthermore, attention mechanism is used to weigh the contribution of each traffic sequence towards the output predictions. The proposed network architecture, when trained end-to-end, results in a superior prediction accuracy compared to baseline models. In addition to contributing towards performance, the attention mechanism creates weight values, which when visualized, provide insights into the decision-making process of the neural network, and consequently produce explainable outputs. Att-MED’s extracted attention weights highlight the contribution of daily and weekly periodic input towards speed prediction.

No abstract available

With the advent of the Internet of Things (IoT), it has become possible to have a variety of data sets generated through numerous types of sensors deployed across large urban areas, thus empowering the notion of smart cities. In smart cities, various types of sensors may fall into different administrative domains and may be accessible through exposed Application Program Interfaces (APIs). In such setups, for traffic prediction in Intelligent Transport Systems (ITS), one of the major prerequisites is the integration of heterogeneous data sources within a preprocessing data pipeline resulting into hybrid feature space. In this paper, we first present a comprehensive algorithm to integrate heterogeneous data obtained from sensors, services, and exogenous data sources into a hybrid spatial–temporal feature space. Following a rigorous exploratory data analysis, we apply a variety of deep learning algorithms specialized for time series geospatial data and perform a comparative analysis of Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Convolutional Neural Network (CNN), and their hybrid combinations. The hybrid LSTM–GRU model outperforms the rest with Root Mean Squared Error (RMSE) of 4.5 and Mean Absolute Percentage Error (MAPE) of 6.67%.

Spatio-temporal models, which combine GNNs (Graph Neural Networks) and RNNs (Recurrent Neural Networks), have shown state-of-the-art accuracy in traffic speed prediction. However, we find that they consider the spatial and temporal dependencies between speeds separately in the two (i.e., space and time) dimensions, thereby unable to exploit the joint-dependencies of speeds in space and time. In this paper, with the evidence via preliminary analysis, we point out the importance of considering individual dependencies between two speeds from all possible points in space and time for accurate traffic speed prediction. Then, we propose an Individual Spatio-Temporal graph (IST-graph) that represents the Individual Spatio-Temporal dependencies (IST-dependencies) very effectively and a Spatio-Temporal Graph ATtention network (ST-GAT), a novel model to predict the future traffic speeds based on the IST-graph and the attention mechanism. The results from our extensive evaluation with five real-world datasets demonstrate (1) the effectiveness of the IST-graph in modeling traffic speed data, (2) the superiority of ST-GAT over 5 state-of-the-art models (i.e., 2-33% gains) in prediction accuracy, and (3) the robustness of our ST-GAT even in abnormal traffic situations.

Traffic speed prediction is an integral part of an intelligent transportation system (ITS) because an advanced knowledge of traffic speed can help taking proactive preventive steps to avoid impending problems and it can also help in trip planning. Traffic speed data comprises a time series that may be modelled using any statistical or machine learning technique. In most of the cases, however, the performance of such models is bottlenecked due to heteroskedasticity usually present in such datasets. ARCH/GARCH family of models are generally used to model variance in such data. This paper presents a novel technique, termed as GARCH‐GRU, based on additive decomposition that splits data into random (residual) and deterministic parts. Random part is normalized using rolling standard deviation. GARCH (1, 1) is used to predict conditional variance of the residual and the predicted variance is then used in the basic model equation along with normalized residual that mimic white noise as required by the model. The data other than residual is modelled using a GRU model. The approach is applied to two datasets corresponding to a downtown road and a motorway. For comparison, the same datasets are exposed to three classical techniques; seasonal ARIMA, CNN and GRU techniques. The results demonstrate that the GARCH‐GRU technique outperforms others for random data of downtown road but fails to handle dynamic variations present in the motorway data.

With the advancement of automatic driving and smart city, it is critical to predict traffic information for traffic management, traffic planning, and traffic safety. When predicting traffic information, the spatial structure of the roads will also affect the traffic flow information, such as speed, occupancy rate, etc. The common method either merely focusing on the temporal feature without the considering the spatial structure, or the method of spatial feature extraction is only applicable to Euclidean structure, which does not apply to Non-Euclidean structure. This paper proposes a traffic speed prediction method based on time classification in combination with spatial Graph Convolutional Network. This method employs Gated Recurrent Unit to extract the temporal correlation and Graph Convolutional Network to extract the traffic network’s spatial structure. In consideration of the varying features of traffic speed on weekdays and weekends in the time dimension, time is divided into two types: weekdays and weekends. Since the structure of the road network will not change in the short term in actual process, the same network structure of spatial graph convolution can reasonably be shared in the spatial dimension after which the two sections are fused for training and prediction. Finally, this proposed method is compared to some baseline models to prove the performance. Generally speaking, this strategy produces more accurate prediction results on the PEMS_BAY and METR_LA data sets than the baseline models.

No abstract available

Traffic prediction is important in applications such as traffic management, route planning, and traffic flow optimization. Traffic speed prediction is an important part of traffic forecasting, which has always been a challenging problem due to the complexity and dynamics of traffic systems. In order to predict traffic speed more accurately, we propose a traffic speed prediction model based on a multi-head attention mechanism and weighted adjacency matrix: MAT-WGCN. MAT-WGCN first uses GCN to extract the road spatial features in the weighted adjacency matrix, and it uses GRU to extract the correlation between speed and time from the original features. Then, the spatial features extracted by GCN and the temporal features extracted by GRU are fused, and a multi-head attention mechanism is introduced to integrate spatiotemporal features, collect and summarize spatiotemporal road information, and realize traffic speed prediction. In this study, the prediction performance of MAT-WGCN was tested on two real datasets, EXPY-TKY and METR-LA, and compared with the performance of traditional methods such as HA and SVR that do not combine spatial features, as well as T-GCN, A3T-GCN, and newer methods such as GCN and NA-DGRU that combine spatial features. The experimental results demonstrate that MAT-WGCN can capture the temporal and spatial characteristics of road conditions, thus enabling accurate traffic speed predictions. Furthermore, the incorporation of a multi-head attention mechanism significantly enhances the robustness of our model.

Accurate traffic prediction is crucial to alleviating traffic congestion in cities. Existing physical sensor-based traffic data acquisition methods have high transmission costs, serious traffic information redundancy, and large calculation volumes for spatiotemporal data processing, thus making it difficult to ensure accuracy and real-time traffic prediction. With the increasing resolution of UAV imagery, the use of unmanned aerial vehicles (UAV) imagery to obtain traffic information has become a hot spot. Still, analyzing and predicting traffic status after extracting traffic information is neglected. We develop a framework for traffic speed extraction and prediction based on UAV imagery processing, which consists of two parts: a traffic information extraction module based on UAV imagery recognition and a traffic speed prediction module based on deep learning. First, we use deep learning methods to automate the extraction of road information, implement vehicle recognition using convolutional neural networks and calculate the average speed of road sections based on panchromatic and multispectral image matching to construct a traffic prediction dataset. Then, we propose an attention-enhanced traffic speed prediction module that considers the spatiotemporal characteristics of traffic data and increases the weights of key roads by extracting important fine-grained spatiotemporal features twice to improve the prediction accuracy of the target roads. Finally, we validate the effectiveness of the proposed method on real data. Compared with the baseline algorithm, our algorithm achieves the best prediction performance regarding accuracy and stability.

A correct prediction of traffic speed can reasonably plan resources in traffic area, avoid congestion and reduce probability of accidents. In this paper, we propose a Speed-Masked Transformer (SMT) model with an improved decoder, which uses Speed-Masked module to mask the traffic speed data. Decoder structure is simplified and speed of historical moments is weighted by an exponential attenuation function. To verify the advantages of SMT model, datasets collected from interstate freeway in California are used to train and validate the prediction model. The SMT model is compared with several other traffic speed prediction models through simulation. Results show that the SMT model performs best in terms of speed prediction for long prediction horizon.

No abstract available

Traffic speed prediction is among the foundations of advanced traffic management and the gradual deployment of internet of things sensors is empowering data-driven approaches for the prediction. Nonetheless, existing research studies mainly focus on short-term traffic prediction that covers up to one hour forecast into the future. Previous long-term prediction approaches experience error accumulation, exposure bias, or generate future data of low granularity. In this paper, a novel data-driven, long-term, high-granularity traffic speed prediction approach is proposed based on recent development of graph deep learning techniques. The proposed model utilizes a predictor-regularizer architecture to embed the spatial-temporal data correlation of traffic dynamics in the prediction process. Graph convolutions are widely adopted in both sub-networks for geometrical latent information extraction and reconstruction. To assess the performance of the proposed approach, comprehensive case studies are conducted on real-world datasets and consistent improvements can be observed over baselines. This work is among the pioneering efforts on network-wide long-term traffic speed prediction. The design principles of the proposed approach can serve as a reference point for future transportation research leveraging deep learning.

Traffic speed prediction is known as an important but challenging problem. In this paper, we propose a novel model, called LC-RNN, to achieve more accurate traffic speed prediction than existing solutions. It takes advantage of both RNN and CNN models by a rational integration of them, so as to learn more meaningful time-series patterns that can adapt to the traffic dynamics of surrounding areas. Furthermore, since traffic evolution is restricted by the underlying road network, a network embedded convolution structure is proposed to capture topology aware features. The fusion with other information, including periodicity and context factors, is also considered to further improve accuracy. Extensive experiments on two real datasets demonstrate that our proposed LC-RNN outperforms six well-known existing methods. 

No abstract available

Timely and accurate traffic speed predictions are an important part of the Intelligent Transportation System (ITS), which provides data support for traffic control and guidance. The speed evolution process is closely related to the topological structure of the road networks and has complex temporal and spatial dependence, in addition to being affected by various external factors. In this study, we propose a new Speed Prediction of Traffic Model Network (SPTMN). The model is largely based on a Temporal Convolution Network (TCN) and a Graph Convolution Network (GCN). The improved TCN is used to complete the extraction of time dimension and local spatial dimension features, and the topological relationship between road nodes is extracted by GCN, to accomplish global spatial dimension feature extraction. Finally, both spatial and temporal features are combined with road parameters to achieve accurate short-term traffic speed predictions. The experimental results show that the SPTMN model obtains the best performance under various road conditions, and compared with eight baseline methods, the prediction error is reduced by at least 8%. Moreover, the SPTMN model has high effectiveness and stability.

Abstract This paper proposes a deep learning based multitask learning (MTL) model to predict network-wide traffic speed, and introduces two methods to improve the prediction performance. The nonlinear Granger causality analysis is used to detect the spatiotemporal causal relationship among various links so as to select the most informative features for the MTL model. Bayesian optimization is employed to tune the hyperparameters of the MTL model with limited computational costs. Numerical experiments are carried out with taxis’ GPS data in an urban road network of Changsha, China, and some conclusions are drawn as follows. The deep learning based MTL model outperforms four deep learning based single task learning (STL) models (i.e., Gated Recurrent Units network, Long Short-term Memory network, Convolutional Gated Recurrent Units network and Temporal Convolutional Network) and three other classic models (i.e., Support Vector Machine, k-Nearest Neighbors and Evolving Fuzzy Neural Network). The nonlinear Granger causality test provides a reliable guide to select the informative features from network-wide links for the MTL model. Compared with two other optimization approaches (i.e., grid search and random search), Bayesian optimization yields a better tuning performance for the MTL model in the prediction accuracy under the budgeted computation cost. In summary, the deep learning based MTL model with nonlinear Granger causality analysis and Bayesian optimization promises the accurate and efficient traffic speed prediction for a large-scale network.

No abstract available

Increasingly serious traffic congestion requires an accurate and timely traffic speed prediction, which will significantly benefit both individual drivers and decision makers in travel planning and traffic management. However, traffic speed prediction is a long-standing and challenging topic. Due to the availability of traffic datasets and powerful computation resources, deep learning becomes a promising solution to this problem. In this paper, based on Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) models, we propose a model named CLM, which is the first to make use of CNN to extract the features of daily and weekly periodicity of traffic speed at the target area and also extract the spatiotemporal features together with the output of CNN by LSTM layers. We conduct comprehensive simulations to assess the performance of our proposed method based on the real-world dataset of Hong Kong. The results indicate that our proposed CLM model can better predict traffic speed in different forecast time periods than the other five competing methods, including SVR, MLP, Lasso, Random forest, and LSTM.

Traffic speed prediction is a significant branch of the intelligent transportation system (ITS). A good prediction could alleviate the non-recurring congestion on the road and provide a strong decision-making basis for traffic management and control. However, it is always a challenging research problem due to the complexity of the road network and the dynamics of traffic conditions. Many deep learning-based methods have been applied to the traffic prediction problem, which could extract both spatial and temporal information efficiently. However, for some dataset that suffers from data paucity problem, the generalization ability of the model is not good and the performance degrades. To tackle this problem, we proposed a novel graph-based generative adversarial network for the traffic speed prediction problem. We design a generative network to generate some fake traffic data and use a discriminative network to distinguish between real and fake targets. The generator consists of a GraphSAGE and LSTM model to learn the representation of spatial-temporal traffic data. Several experiments have been conducted on several real-world traffic datasets, demonstrating that our proposed model outperforms other baseline models. The experiment results illustrate the importance of utilizing GAN in the training process, which improves the generalization ability of the prediction model.

Traffic speed prediction is essential for efficient traffic operation and management by distributing demand concentration in time and space. To make an accurate prediction, it is required to consider spatio-temporal characteristics of the traffic evolution. Recently, deep learning-based approaches, especially Graph Neural Network (GNN) has been widely adopted to reflect the stated characteristics. However, existing GNN models mainly used for short-term prediction, whereas long-term traffic prediction is more useful by enabling earlier and efficient decisions of traffic management as well as individual travels. In this study, we propose Asymmetric Long-Term Graph Multi-Attention Network (ALT-GMAN) algorithm, an extension of the GMAN. ALT-GMAN can predict short and long-term traffic speed by considering asymmetric characteristics of forward and backward waves observed in real roadways. ALT-GMAN is tested with six months highway data of PeMS-Bay area, and MAPE for 3-hours and 6-hours prediction is evaluated as 5.53% and 6.05%, respectively. ALT-GMAN outperforms the existing models in short-term speed prediction, and provides a robust performance in long-term prediction problems, too.

Traffic speed prediction plays a fundamental role in traffic management and driving route planning. However, timely accurate traffic speed prediction is challenging as it is affected by complex spatial and temporal correlations. Most existing works cannot simultaneously model spatial and temporal correlations in traffic data, resulting in unsatisfactory prediction performance. In this article, we propose a novel hybrid deep learning approach, named HDL4TSP, to predict traffic speed in each region of a city, which consists of an input layer, a spatial layer, a temporal layer, a fusion layer, and an output layer. Specifically, first, the spatial layer employs graph convolutional networks to capture spatial near dependencies and spatial distant dependencies in the spatial dimension. Second, the temporal layer employs convolutional long short-term memory (ConvLSTM) networks to model closeness, daily periodicity, and weekly periodicity in the temporal dimension. Third, the fusion layer designs a fusion component to merge the outputs of ConvLSTM networks. Finally, we conduct extensive experiments and experimental results to show that HDL4TSP outperforms four baselines on two real-world data sets.

Traffic speed prediction is an incredibly important subject of Intelligent transportation system (ITS). Efficient speed prediction methods greatly contribute to reducing traffic congestion. Most existing models focus on short term while the long-term speed prediction for a whole day is not completely developed. In this paper, a Geometric Algebra Convolutional LSTM and Graph Attention (GAConvLSTM-GAT) model is proposed to raise a potential for achieving long-term speed prediction. The proposed model is composed of a Geometric Algebra ConvLSTM (GAConvLSTM) module to extract the spatial-temporal feature, and a Graph Attention (GAT) module to make speed predictions based on the features. The experiments are evaluated by two elevated highway traffic datasets. The presented results illustrate that our GAConvLSTM model outperforms multiple baseline neural network methods.

The difficulty of dealing with traffic jams, pollution, road accidents, and any other disturbances in the management of the city becomes more and more troublesome as the traffic increases. So, adequate traffic management is required. So, our study includes traffic prediction for particular weather using machine learning and deep learning techniques, including Random Forest (RF), Long Short Term Memory (LSTM), AutoEncoders, and Generative Adversarial Networks (GAN). The research highlights the utility of such models in forecasting traffic patterns and creating realistic synthetic data for simulation by analyzing the static and temporal aspects of the traffic data. The results show that these systems enhance traffic management systems and facilitate the development of smarter cities.

No abstract available

A two‐step framework that integrates real‐time data collection with time series forecasting models for predicting traffic volume is proposed. In the first step, the framework utilizes live highway surveillance video data and YOLO‐v7 object detector to construct accurate traffic volume data. In the second step, an ARIMA–LSTM time series model is applied to forecast future traffic volumes. Experimental results show that YOLO‐v7 achieved a vehicle detection accuracy of over 93.30%, ensuring high precision in traffic volume data construction. The ARIMA–LSTM model demonstrated superior performance in traffic volume prediction, with a mean squared error of 87.97, root mean squared error of 10,388.57, and mean absolute error of 101.39. YOLO‐v7's detection speed of 7.8 ms per frame further validates the feasibility of real‐time data construction. The findings indicate that the combination of YOLO‐v7 for vehicle detection and ARIMA–LSTM for traffic prediction is highly effective, offering a significant reduction in training time compared to more complex deep learning models while maintaining high prediction accuracy. This research presents a unified solution for traffic data collection and prediction, enhancing transportation infrastructure planning and optimizing traffic flow. Future work will focus on extending the prediction intervals and further refining the models to improve performance.

Advanced traffic management systems rely heavily on accurate traffic state estimation and prediction. Traffic prediction based on conventional road-based sensors faces considerable challenges due to spatiotemporal correlations of traffic flow propagation, and heterogeneous, error-prone, and missing data. This paper proposes a hybrid deep learning approach for online traffic volume prediction in an urban network. The approach ensembles the long short-term memory (LSTM) neural network and the convolutional neural networks (CNN) in a parallel way. In order to deal with missing data, a state-of-the-art Bayesian probabilistic imputation method is employed in the overall prediction pipeline. The hybrid traffic prediction structure can capture the spatiotemporal characteristics of traffic volume. The proposed prediction model is verified by the loop and infrared sensor data in the inner city network of the City of Dresden. Experimental results show that it can achieve superior volume prediction compared with baseline methods.

Hourly traffic volume prediction is now emerging to mitigate and respond to hourly-level traffic congestion augmented by deep learning techniques. Incorporating meteorological data into the forecasting of hourly traffic volumes substantively improves the precision of long-term traffic forecasts. Nonetheless, integrating weather data into traffic prediction models is challenging due to the complex interplay between traffic flow, time-based patterns, and meteorological conditions. This paper proposes a graph convolutional network to predict long-term traffic volume with meteorological information. This study utilized a four-year traffic volume and meteorological information dataset in Chung-ju si to train and validate the models. The proposed model performed better than the other baseline scenarios with conventional and state-of-the-art deep learning techniques. Furthermore, the counterfactual scenarios analysis revealed the potential negative impacts of meteorological conditions on traffic volume. These findings will enable transportation planners predict hourly traffic volumes for different scenarios, such as harsh weather conditions or holidays. Furthermore, predicting the microscopic traffic simulation for different scenarios of weather conditions or holidays is useful.

Real-time traffic volume prediction is critical for proactive traffic control and road guidance via fast-communication networks. However, existing research relies on historical traffic volume and fails to consider vehicle travel route choices, leading to lower prediction accuracy and insufficient capture of real-time travel dynamics. To address this limitation, we propose a novel approach for real-time traffic volume prediction on urban links that incorporates vehicle travel trajectories extracted from recorded license plate recognition data using Radio Frequency Identification. A Recurrent Neural Network is used to identify route patterns and predict future routes, and link traffic volume is estimated by aggregating all predicted vehicle trajectories that pass through each base station. Our approach achieves a high prediction accuracy with a mean value of 85.19% and a standard deviation of 5.29%. Comparing our approach with a traditional Recurrent Neural Network based on historical traffic volume data, the vehicle trajectory-based approach yields an average accuracy improvement of 39.70%. These results highlight the superiority of our approach for predicting traffic volume in real-time and demonstrate its potential as a valuable tool for traffic control and road guidance.

With the advancement of artificial intelligence, traffic forecasting is gaining more and more interest in optimizing route planning and enhancing service quality. Traffic volume is an influential parameter for planning and operating traffic structures. This study proposed an improved ensemble-based deep learning method to solve traffic volume prediction problems. A set of optimal hyperparameters is also applied for the suggested approach to improve the performance of the learning process. The fusion of these methodologies aims to harness ensemble empirical mode decomposition’s capacity to discern complex traffic patterns and long short-term memory’s proficiency in learning temporal relationships. Firstly, a dataset for automatic vehicle identification is obtained and utilized in the preprocessing stage of the ensemble empirical mode decomposition model. The second aspect involves predicting traffic volume using the long short-term memory algorithm. Next, the study employs a trial-and-error approach to select a set of optimal hyperparameters, including the lookback window, the number of neurons in the hidden layers, and the gradient descent optimization. Finally, the fusion of the obtained results leads to a final traffic volume prediction. The experimental results show that the proposed method outperforms other benchmarks regarding various evaluation measures, including mean absolute error, root mean squared error, mean absolute percentage error, and R-squared. The achieved R-squared value reaches an impressive 98%, while the other evaluation indices surpass the competing. These findings highlight the accuracy of traffic pattern prediction. Consequently, this offers promising prospects for enhancing transportation management systems and urban infrastructure planning.

No abstract available

Predicting urban traffic volume presents significant challenges due to complex temporal dependencies and fluctuations driven by environmental and situational factors. This study addresses these challenges by evaluating the effectiveness of three deep learning architectures— Recurrent Neural Network (RNN), Gated Recurrent Unit (GRU), and Convolutional Neural Network (CNN)—in forecasting hourly traffic volume on Interstate 94. Using a standardized dataset, each model was assessed on predictive accuracy, computational efficiency, and suitability for real-time applications, with Mean Absolute Percentage Error (MAPE), Root Mean Square Error (RMSE), R2 coefficient, and computation time as performance metrics. The GRU model demonstrated the highest accuracy, achieving a MAPE of 2.105%, RMSE of 0.198, and R2 of 0.469, but incurred the longest computation time of 7917 seconds. Conversely, CNN achieved the fastest computation time at 853 seconds, with moderate accuracy (MAPE of 2.492%, RMSE of 0.214, R2 of 0.384), indicating its suitability for real- time deployment. The RNN model exhibited intermediate performance, with a MAPE of 2.654% and RMSE of 0.215, reflecting its limitations in capturing long-term dependencies. These findings highlight crucial trade- offs between accuracy and efficiency, underscoring the need for model selection aligned with specific application requirements. Future work will explore hybrid architectures and optimization strategies to enhance further predictive accuracy and computational feasibility for urban traffic management.

No abstract available

In response to the problem of fixed time intervals for short-term traffic flow prediction, which fails to meet the requirements of traffic signal control based on traffic cycle signals, this paper proposes an improved long short-term memory-based method for periodic traffic volume prediction. The method presented in this study involves improvements to the Long Short-Term Memory (iLSTM) and Bidirectional Long Short-Term Memory (iBiLSTM) models, leading to the construction of the iBiLSTM-iLSTM-NN model. This model incorporates spatial data from surrounding intersections and employs data fitting techniques to establish the correlation between periodic queue length and traffic volume. Subsequently, a predictive model for periodic traffic volume is developed based on this correlation, enabling reliable forecasting of future traffic volumes within a given cycle. Additionally, actual intersection data is collected for simulation analysis. The results indicate that the prediction error of periodic traffic volume is influenced by different traffic flow characteristics such as peak, off-peak, and normal periods, as well as different inbound lanes. Different model parameters have a noticeable impact on the model’s performance, with smaller batch sizes leading to more stable models. By comparing the performance of different prediction models using various error evaluation metrics, this study finds that the proposed model exhibits the most stable performance. The research findings can be applied to rapidly predict future traffic volumes for several periods based on the instantaneous queue length at the end of the red signal phase, providing reliable, accurate, and timely data for urban traffic signal control.

Accurate traffic volume prediction plays a crucial role in urban traffic control by relieving congestion through improved regulation of traffic volume. Network‐level traffic volume prediction and detector failure have rarely been considered in the literature. This paper proposes a framework based on long short‐term memory and the multilayer perceptron that can predict network‐level traffic volumes even with detector failure. A profile model learns the profile of the detector's signature (traffic pattern). Detectors with similar profiles are considered to have similar traffic patterns and are grouped into a cluster. Failed detectors can obtain reference information from similar detectors in the same cluster without additional information. A predictive model is developed for each cluster. The proposed method is validated using Japan Road Traffic Information Center data for three cities. The computational results indicate that the proposed method performs well both on typical days and atypical days (the COVID‐19 lockdown period and the 2021 Tokyo Olympics). Further, it considers detector reliability: the increase in mean absolute error is less than 1 veh/5 min when the probability of detector failure increases to 20%.

No abstract available

No abstract available

The rapid growth of industry and the economy has contributed to a tremendous increase in traffic in all urban areas. People face the problem of traffic congestion frequently in their day-to-day life. To alleviate congestion and provide traffic guidance and control, several types of research have been carried out in the past to develop suitable computational models for short- and long-term traffic. This study developed an effective multi-dimensional dataset-based model in cyber–physical systems for more accurate traffic-volume prediction. The integration of quantum convolutional neural network and Bayesian optimization (QCNN_BaOpt) constituted the proposed model in this study. Furthermore, optimal tuning of hyperparameters was carried out using Bayesian optimization. The constructed model was evaluated using the US accident dataset records available in Kaggle, which comprise 1.5 million records. The dataset consists of 47 attributes describing spatial and temporal behavior, accidents, and weather characteristics. The efficiency of the proposed model was evaluated by calculating various metrics. The performance of the proposed model was assessed as having an accuracy of 99.3%. Furthermore, the proposed model was compared against the existing state-of-the-art models to demonstrate its superiority.

No abstract available

No abstract available

No abstract available

The purpose of this paper is to investigate the factors affecting the passenger traffic volume on the high-capacity MRT, to analyze the traffic volume trends, and to make traffic volume predictions more accurately. This paper uses the LSTM (Long Short-term Memory) model and regression analysis for Taipei MRT passenger traffic volume prediction. The training data, including dates, temperatures and number of passengers, were collected, and a regression analysis was performed to find trends in the data.

Many countries around the globe have imposed several response measures to suppress the rapid spread of the COVID-19 pandemic since the beginning of 2020. These measures have impacted routine daily activities, along with their impact on economy, education, social and recreational activities, and domestic and international travels. Intuitively, the different imposed policies and measures have indirect impacts on urban traffic mobility. As a result of those imposed measures and policies, urban traffic flows have changed. However, those impacts are neither measured nor quantified. Therefore, estimating the impact of these combined yet different policies and measures on urban traffic flows is a challenging task. This paper demonstrates the development of an artificial neural networks (ANN) model which correlates the impact of the imposed response measure and other factors on urban traffic flows. The results show that the adopted ANN model is capable of mapping the complex relationship between traffic flows and the response measures with a high level of accuracy and good performance. The predicted values are closed to the observed ones. They are clustered around the regression line, with a coefficient of determination (R2) of 0.9761. Furthermore, the developed model can be generalized to determine the anticipated demand levels resulted from imposing any of the response measures in the post-pandemic era. This model can be used to manage traffic during mega-events. It can be also utilized for disaster or emergency situations, where traffic flow estimates are highly required for operational and planning purposes.

Traffic prediction for scenic spots is an important topic in modeling an urban traffic system. Existing traffic prediction approaches typically use raw traffic data and road networks without considering the physical environment and human–environment interaction. This article presents a novel traffic prediction model that considers: (1) the topological structure of the city road network; (2) the popularity and accessibility of each scenic spot in the city; and (3) the traffic volumes of nearby scenic spots. The proposed model first learns a series of traffic dependency graphs by the Multi‐graph Convolutional Network using multiple data sources describing historical traffic volumes, scenic spots popularity, land function, location, and accessibility. The graph nodes represent the scenic spots, and the links between them represent their traffic dependency, considering all traffic and geographic features. Then the proposed model uses the Gated Recurrent Unit (GRU) to capture the temporal dependency between multiple fused graphs for traffic volume prediction. The experiments show that the proposed model (M‐GCNGRU) can effectively exploit and integrate geographic data with historical traffic data for traffic volume prediction, outperforming several classical and state‐of‐the‐art methods.

Predicting short-term traffic volume is essential to improve transportation systems management and operations (TSM0) and the overall efficiency of traffic networks. The real-time data, collected from Internet of Things (loT) devices, can be used to predict traffic volume. More specifically, the Automated Traffic Signal Performance Measures (ATSPM) data contain high-fidelity traffic data at multiple intersections and can reveal the spatio-temporal patterns of traffic volume for each signal. In this study, we have developed a machine learning-based approach using the data collected from ATSPM sensors of a corridor in Orlando, FL to predict future hourly traffic. The hourly predictions are calculated based on the previous six hours volume seen at the selected intersections. Additional factors that play an important role in traffic fluctuations include peak hours, day of the week, holidays, among others. Multiple machine learning models are applied to the dataset to determine the model with the best performance. Random Forest, XGBoost, and LSTM models show the best performance in predicting hourly traffic volumes.

The growth in the transportation sector has led to an enormous increase in the number of vehicles that ply our roads daily. Even though this advancement has provided numerous transportation modes, it has resulted in serious transportation issues including road congestion. Hence, estimating the number of vehicles on a road will enable traffic managers to take appropriate decisions to curb congestion. In this paper, we propose to use an extreme gradient boosting (XGBoost) algorithm to efficiently and accurately predict the hourly traffic volume. We investigate the effectiveness of the proposed method for different scenarios including how well it performs during extreme weather conditions and holidays. We further investigate the effect of ridge and LASSO regularization on the performance of XGBoost. We then propose a new approach for setting the LASSO regularization parameter in terms of the number of observations and predictors. The performance and computational efficiency of the proposed approach is evaluated on data collected from Interstate-94, Minnesota and the results are compared with existing methods. The results show that the proposed method provides a good balance between performance and computational efficiency.

No abstract available

Abstract As a potential solution to relieve traffic congestions and help build a more safe traffic system, traffic flow prediction methods are given much attention in recent years. In previous studies, it can be found the machine learning (ML)-based methods are widely used in volume predictions of single roads. However, when applied in a more complicated road network, they usually show low efficiency and need to pay higher computing costs. To solve this problem, an innovative ML-based model, named Selected Stacked Gated Recurrent Units model (SSGRU), is proposed in this paper, which is mainly in allusion to road network traffic flow. There are mainly two parts in this model: one is used to do spatial pattern mining based on linear regression coefficients, and the other one includes a stacked gated recurrent unit (SGRU), which is essential for multi-road traffic flow prediction. As the basic unit, a simple tree structure is adopted to approximate the given road network. Particularly, we implemented our model into both suburban and urban traffic contexts, to prove its high adaptability. The whole evaluation process is based on seven different traffic volume data sets recorded at the 15-min interval, chosen from the England Highways. The results show that our model has higher accuracy than others when applied to a multi-road input infrastructure for all road scenarios.

Knowledge of the truck traffic volumes on state and interstate highways is critical for highway authorities and federal organizations. Increased urbanization, population growth, and economic development have led to an increased demand for freight travel. Several planning applications demand reliable and accurate truck traffic prediction. A review of the available literature indicated that limited research had been performed on the development and utilization of a universal automatic framework for truck traffic volume prediction. As a result, there is a gap to incorporate inclusive predictors, a broad dataset, a comprehensive feature selection approach, and a robust cross-validation method that utilizes both linear and non-linear algorithms. The present study uses a hyperparameter optimization framework to select the appropriate feature selection method and modeling approach among a comprehensive list of available state of the art approaches. Distinct from models based on individual case studies, the proposed framework allows for greater customization and minimized MAPE error. The developed framework automates much of the traffic count forecasting process, and the resulting method is less labor-intensive and may be utilized without the need for experienced data analysts. Florida’s interstate highways historical traffic data were used to test the feasibility of the proposed framework. The results of the Florida Case Study revealed the superiority of non-linear models in the generalization and prediction of traffic volumes over linear models. The random forest algorithm results on the test dataset in this study demonstrate this model’s ability to predict truck traffic with 86% accuracy. Spatial variables were the most significant variable group, followed by road characteristics.

Long short-term memory (LSTM) neural network shows excellent performance in learning, processing, and classifying time series data but with some limitations such as high computational cost and lack of interpretability. Fuzzy neural networks, which combine the complementary capabilities of both neural networks and fuzzy system, thus, constitute a more promising technique for processing traffic flow. This article presents a Type-2 fuzzy LSTM (T2F-LSTM) neural network model for long-term traffic volume prediction. T2F Sets (T2FSs) provide more freedom to describe membership information and process data with higher uncertainty better than the traditional fuzzy system does. In this article, an interval T2FSs is introduced to extract the probability distribution and spatial–temporal characteristics of traffic volume. Using parameters of the closure of support obtained in interval T2FSs, weights of input gate in LSTM neural network are updated and converged to the region with a larger slope of the sigmoid function fast. The network interpretability is also increased by better control of the information flow using motivational factors constructed from the parameters. Experiment conducted with real traffic volume data shows that the proposed model achieves more accurate prediction results and shorter network training time.

The prediction of short-term volatile traffic becomes increasingly critical for efficient traffic engineering in intelligent transportation systems. Accurate forecast results can assist in traffic management and pedestrian route selection, which will help alleviate the huge congestion problem in the system. This paper presents a novel hybrid DTMGP model to accurately forecast the volume of passenger flows multi-step ahead with the comprehensive consideration of factors from temporal, origin-destination spatial, and frequency and self-similarity perspectives. We first apply discrete wavelet transform to decompose the traffic volume series into an appropriation component and several detailed components. Then we propose a more efficient tracking model to forecast the appropriation component and a novel Gaussian process model to forecast the detailed components. The forecasting performance is evaluated with real-time passenger flow data in Chongqing, China. Simulation results demonstrate that our hybrid model can achieve on average 20%–50% accuracy improvement, especially during rush hours.

No abstract available

Geostatistical methods have been widely used for spatial prediction and the assessment of traffic issues. Most previous studies use point-based interpolation, but they ignore the critical information of the road segment itself. This can lead to inaccurate predictions, which will negatively affect decision making of road agencies. To address this problem, segment-based regression kriging (SRK) is proposed for traffic volume prediction with differentiation between heavy and light vehicles in the Wheatbelt region of Western Australia. Cross validations reveal that the prediction accuracy for heavy vehicles is significantly improved by SRK (R2 = 0.677). Specifically, 78% of spatial variance and 53% of estimated uncertainty are improved by SRK for heavy vehicles compared with regression kriging, a best performing point-based geostatistical model. This improvement shows that SRK can provide new insights into the spatial characteristics and spatial homogeneity of a road segment. Implementation results of SRK-based predictions show that the impact of heavy vehicles on road maintenance is much larger than that of light vehicles and it varies across space, and the total impacts of heavy vehicles account for more than 82% of the road maintenance burden even though its volume only accounts for 21% of traffic.

Abstract Because of the rapid changes in traffic conditions caused by various circumstances, such as road construction and traffic jams, the distribution of the traffic volume data changes over time. The performances of traditional traffic volume prediction methods, with fixed model types and parameter settings, suffer from gradual degradation during these concept drift processes. In this paper, a novel incremental regression framework under the concept drifting environment is proposed, with ensemble learning as the major solution for updating the distribution representation. First, we transform the regression problem of traffic volume forecasting into a binary classification problem. Second, loss functions for incremental and ensemble learning are constructed based on this transformation. Finally, the incremental learning of the regression function is formulated as stepwise updating of the decision hyperplane. The experimental results show that our method is more stable and accurate than the existing incremental and ensemble regression methods.

This paper uses Gaussian interval type-2 fuzzy set theory on historical traffic volume data processing to obtain a 24-hour prediction of traffic volume with high precision. A K-means clustering method is used in this paper to get 5 minutes traffic volume variation as input data for the Gaussian interval type-2 fuzzy sets which can reflect the distribution of historical traffic volume in one statistical period. Moreover, the cluster with the largest collection of data obtained by K-means clustering method is calculated to get the key parameters of type-2 fuzzy sets, mean and standard deviation of the Gaussian membership function. Using the range of data as the input of Gaussian interval type-2 fuzzy sets leads to the range of traffic volume forecasting output with the ability of describing the possible range of the traffic volume as well as the traffic volume prediction data with high accuracy. The simulation results show that the average relative error is reduced to 8% based on the combined K-means Gaussian interval type-2 fuzzy sets forecasting method. The fluctuation range in terms of an upper and a lower forecasting traffic volume completely envelopes the actual traffic volume and reproduces the fluctuation range of traffic flow.

How to provide accurate and timely traffic flow information has become a hot topic in recent years since they can help schedule trips better and reduce traffic congestion. In previous studies, some machine learning (ML)-based models were proposed to predict the traffic volume at a single road segment/position, and these models performed not bad. However, when applied in a more complicated road network, they show low efficiency or need to pay higher computing costs. To solve this problem, an innovative ML-based model named selected stacked gated recurrent units model (SSGRU), is proposed for predicting the traffic flow through a sparse road network in this paper. There are mainly two parts in this model, one is used to do spatial pattern mining based on linear regression coefficients, and the other one includes a stacked gated recurrent unit (SGRU) which is essential for multi-road traffic flow prediction. A binarytree is adopted to approximate the sparse road network in the suburban area. To evaluate the proposed model, seven different traffic volume data sets recorded at 15-min interval are chosen from the England Highways data set to test our proposed work. The result shows that our model has greater adaptability and higher accuracy than others when applied to a multi-road input infrastructure.

Reliable traffic flow prediction can greatly support the Intelligent Transportation System (ITS) to generate more effective traffic management decisions. Previous volume predictions mainly focused on the single road with simple flow patterns, such as suburban highways. However, with the development of the urban transportation system, the reliable flow information support becomes more significant for forming a solid ITS. Besides, travel delay is another widely neglected problem but can affect the prediction result significantly. Specifically, vehicles need some time to move from one place to another, and this time is called the travel delay. For further enhancing the prediction performance under the urban scenario, we propose a delay-based deep learning framework (MDGRU) to improve the accuracy of the short-term traffic flow prediction, in which travel delay is handled in the form of a weighted matrix enrolled into a multivariate input stacked Recurrent Neural Network (RNN). Multivariate input makes this approach has a stronger mining ability for spatial relationships capture, and the stacked structure leads to a more accurate pattern learning process. The results show that our approach is accurate and reliable.

No abstract available

This study evaluates the use of the Artificial Bee Colony (ABC) algorithm to optimize the Recurrent Neural Network (RNN) that is used to analyze traffic volume. Related studies have shown that Deep Neural Networks are superseding the Shallow Neural Networks especially in terms of performance. Here we show that using the ABC algorithm in training the Recurrent Neural Network yields better results, compared to several other algorithms that are based on statistical or heuristic techniques that were preferred in earlier studies. The ABC algorithm is an example of swarm intelligence algorithms which are inspired by nature. Therefore, this study evaluates the performance of the RNN trained using the ABC algorithm for the purpose of forecasting. The performance metric used in this study is the Mean Squared Error (MSE) and ultimately, the outcome of the study may be generalized and extended to suit other domains.

In recent years, the transformer architecture and its variants have been widely applied for multivariate time series forecasting. They have achieved state-of-the-art performance across many domains, e.g. road traffic prediction, weather forecasting, exchange rates forecasting etc. This success is based on their long sequence modelling ability. The self-attention mechanism, the core of the standard transformer architecture and its variants, is position invariant. This is typically dealt with by incorporating positional encoding in the input. However, this alone is insufficient in the case of time series forecasting where the relative position of occurrence of various features is crucial. In this work we propose a new transformer architecture for traffic state forecasting that is more position sensitive. More specifically, we build position awareness into the attention weighted value computation process. We refer to our model as the Position Aware Transformer (PATr). Our experiments validate the performance of PATr extensively using the well-known PeMS traffic dataset. Our results show appreciable improvement in performance compared to the previous state-of-the-art.

Traffic-state forecasting is crucial for traffic management and control strategies, as well as user- and system-level decision-making in the transportation network. While traffic forecasting has been approached with a variety of techniques over the last couple of decades, most approaches simply rely on endogenous traffic variables for state prediction, despite the evidence that exogenous factors can significantly affect traffic conditions. This paper proposes a multidimensional spatiotemporal graph attention-based traffic-prediction approach (M-STGAT), which predicts traffic based on past observations of speed, along with lane-closure events, temperature, and visibility across a large transportation network. The approach is based on a graph attention network architecture, which learns based on the structure of the transportation network on which these variables are observed. Numerical experiments are performed using traffic-speed and lane-closure data from the Caltrans Performance Measurement System (PeMS) and corresponding weather data from the National Oceanic and Atmospheric Administration (NOOA) Automated Surface Observing Systems (ASOS). The numerical experiments implement three alternative models which do not allow for multidimensional input, along with two alternative multidimensional models, based on the literature. The M-STGAT outperforms the five alternative models in validation and testing with the primary data set, as well as for one transfer data set across all three prediction horizons for all error measures. However, the model’s transferability varies for the remaining two transfer data sets, which may require further investigation. The results demonstrate that M-STGAT has the most consistently low error values across all transfer data sets and prediction horizons.

No abstract available

Short-term traffic state forecasting is critical for real-time traffic control, but due to its complexity and its nonlinear nature, it is difficult to obtain a high degree of precision. The “k-nearest neighbors” model has been widely used to solve nonlinear regression and time series forecasting. This paper presents a traffic state forecasting method using adaptive neighborhood selection based on expansion strategy to search manifold neighbors to get higher precision with manifold neighbors. We propose a method of linear structure to handle the traffic data in Euclidean space to find a manifold neighbor that is more suitable for predicting traffic states. The results of extensive comparison experiments indicate that the proposed model can produce more accurate forecasting results than other classic algorithms.

No abstract available

Timely accurate traffic forecast is crucial for urban traffic control and guidance. Due to the high nonlinearity and complexity of traffic flow, traditional methods cannot satisfy the requirements of mid-and-long term prediction tasks and often neglect spatial and temporal dependencies. In this paper, we propose a novel deep learning framework, Spatio-Temporal Graph Convolutional Networks (STGCN), to tackle the time series prediction problem in traffic domain. Instead of applying regular convolutional and recurrent units, we formulate the problem on graphs and build the model with complete convolutional structures, which enable much faster training speed with fewer parameters. Experiments show that our model STGCN effectively captures comprehensive spatio-temporal correlations through modeling multi-scale traffic networks and consistently outperforms state-of-the-art baselines on various real-world traffic datasets.

Short-term traffic forecasting based on deep learning methods, especially recurrent neural networks (RNN), has received much attention in recent years. However, the potential of RNN-based models in traffic forecasting has not yet been fully exploited in terms of the predictive power of spatial-temporal data and the capability of handling missing data. In this paper, we focus on RNN-based models and attempt to reformulate the way to incorporate RNN and its variants into traffic prediction models. A stacked bidirectional and unidirectional LSTM network architecture (SBU-LSTM) is proposed to assist the design of neural network structures for traffic state forecasting. As a key component of the architecture, the bidirectional LSTM (BDLSM) is exploited to capture the forward and backward temporal dependencies in spatiotemporal data. To deal with missing values in spatial-temporal data, we also propose a data imputation mechanism in the LSTM structure (LSTM-I) by designing an imputation unit to infer missing values and assist traffic prediction. The bidirectional version of LSTM-I is incorporated in the SBU-LSTM architecture. Two real-world network-wide traffic state datasets are used to conduct experiments and published to facilitate further traffic prediction research. The prediction performance of multiple types of multi-layer LSTM or BDLSTM models is evaluated. Experimental results indicate that the proposed SBU-LSTM architecture, especially the two-layer BDLSTM network, can achieve superior performance for the network-wide traffic prediction in both accuracy and robustness. Further, comprehensive comparison results show that the proposed data imputation mechanism in the RNN-based models can achieve outstanding prediction performance when the model's input data contains different patterns of missing values.

No abstract available

With the rapid expansion of smart cities, intelligent transportation systems (ITSs) are assuming an increasingly pivotal role. Among the multitude of tasks within ITS, traffic state forecasting stands out. As city boundaries grow, traffic forecasting encounters scalability and network transmission challenges. This research contributes to traffic state forecasting within large-scale, massive Internet of Things (IoT) scenarios. By investigating an urban expressway managing architecture that employs long range wide area network (LoRaWAN) communication, a novel deep learning-based model named Time Alignment-based temporal-graph attention network (TATGaN) is proposed. Using the temporal-graph attention mechanism, TATGaN is able to extract temporal-spatial information and predict traffic state accurately. Moreover, the time alignment block makes TATGaN capable of handling irregular sequences given by the asynchronous arrival of packets. Simulation results based on OpenStreetMap data of a specific region within Abu Dhabi show that TATGaN outperforms existing baseline methods in prediction performance with lower error and the higher reliability. Furthermore, the performance evaluation demonstrates the suitability of TATGaN for large-scale traffic network scenarios, attributing its efficiency to the transmission schedule and parameter mechanisms in LoRaWAN networks.

Traffic state estimation is an essential component of Intelligent Transportation System (ITS) designed for alleviating traffic congestion. As traffic data is composed of intricate information which can also be impacted by various factors, scholars have been attempting to utilize state-of-the-art deep learning forecasting models in recent years. However, a more complex and robust model is required to extract long-range correlations with large-scale traffic data sequences. In order to overcome the weaknesses of deep learning models, the superior performance of transformers is expected to address this effectively in time-series forecasting with transport data. Employing the capabilities of transformers in extracting long-term trends and dynamic dependencies, the proposed model improves the deep learning prediction performance for real datasets. The findings indicate that the transformer-based model exhibited promising performance in forecasting long-term traffic patterns and characteristics with a large quantity of data. In this paper, a comparison across conventional hybrid deep learning models with the Spatio-Temporal Autoencoder Transformer (STAT) model was conducted using real-world datasets. The multi-head attention-based transformer model outperformed all other comparative approaches for large-scale data demonstrating its importance in measuring the error criteria. Comprehensive evaluations across various traffic prediction datasets have established the validity of the proposed approach. Also, for an efficient model selection, Akaike Information Criterion (AIC), Schwarz Bayesian Information Criterion (SBIC), Hannan-Quinn Information Criterion (HQIC), and corrected AIC (AICc) tools were used to evaluate and compare models based on their ability to balance fit and complexity. The proposed model has shown improvements in RMSE for the STREET I980 and PEMSBAY datasets of 1.03 and 0.27, respectively. In RMSE for 60 minutes, for the same datasets, the improvements were 0.84 for STREET I980 and 0.56 for PEMS-BAY. These findings underscore the potential of transformer-based models in enhancing the performance of traffic prediction systems.

Long-sequence traffic flow forecasting plays a crucial role in intelligent transportation systems. However, existing Transformer-based approaches face a quadratic complexity bottleneck in computation and are prone to over-smoothing in deep architectures. This results in overly averaged predictions that fail to capture the peaks and troughs of traffic flow. To address these issues, we propose a State-Space Generative Adversarial Network (SSGAN) with a state-space generator and a multi-scale convolutional discriminator. Specifically, a bidirectional Mamba-2 model was designed as the generator to leverage the linear complexity and efficient forecasting capability of state-space models for long-sequence modeling. Meanwhile, the discriminator incorporates a multi-scale convolutional structure to extract traffic features from the frequency domain, thereby capturing flow patterns across different scales, alleviating the over-smoothing issue and enhancing discriminative ability. Through adversarial training, the model is able to better approximate the true distribution of traffic flow. Experiments conducted on four real-world public traffic flow datasets demonstrate that the proposed method outperformed the baselines in both forecasting accuracy and computational efficiency.

The study on the spatial-temporal characteristics of highway traffic flow is helpful to deeply understand the inherent evolution of highway traffic system and provide a theoretical basis for prediction and control of highway traffic flow. This paper makes an empirical analysis on the spatial-temporal characteristics of highway traffic flow using manifold similarity index and manifold learning technology. The time series of highway traffic flow is converted into the distance series containing manifold features to calculate the manifold distance between multi-section traffic flow data points, which are highly similar to spatial-temporal distribution of traffic flow speed parameters, and then, the levels calibration of traffic state is carried out according to the manifold distance, so as to reveal the distribution rule of spatial-temporal characteristics of highway traffic flow. Its prediction error is obviously lower than the traditional distance measurement method, which has higher accuracy. The research of this paper can provide new ideas and methods to reveal the highway traffic flow evolution and traffic state prediction.

With the rapid development of the Intelligent Transportation System (ITS), accurate traffic forecasting has emerged as a critical challenge. The key bottleneck lies in capturing the intricate spatio-temporal traffic patterns. In recent years, numerous neural networks with complicated architectures have been proposed to address this issue. However, the advancements in network architectures have encountered diminishing performance gains. In this study, we present a novel component called spatio-temporal adaptive embedding that can yield outstanding results with vanilla transformers. Our proposed Spatio-Temporal Adaptive Embedding transformer (STAEformer) achieves state-of-the-art performance on five real-world traffic forecasting datasets. Further experiments demonstrate that spatio-temporal adaptive embedding plays a crucial role in traffic forecasting by effectively capturing intrinsic spatio-temporal relations and chronological information in traffic time series.

This paper investigates traffic forecasting, which attempts to forecast the future state of traffic based on historical situations. This problem has received ever-increasing attention in various scenarios and facilitated the development of numerous downstream applications such as urban planning and transportation management. However, the efficacy of existing methods remains sub-optimal due to their tendency to model temporal and spatial relationships independently, thereby inadequately accounting for complex high-order interactions of both worlds. Moreover, the diversity of transitional patterns in traffic forecasting makes them challenging to capture for existing approaches, warranting a deeper exploration of their diversity. Toward this end, this paper proposes Conjoint Spatio-Temporal graph neural network (abbreviated as COOL), which models heterogeneous graphs from prior and posterior information to conjointly capture high-order spatio-temporal relationships. On the one hand, heterogeneous graphs connecting sequential observation are constructed to extract composite spatio-temporal relationships via prior message passing. On the other hand, we model dynamic relationships using constructed affinity and penalty graphs, which guide posterior message passing to incorporate complementary semantic information into node representations. Moreover, to capture diverse transitional properties to enhance traffic forecasting, we propose a conjoint self-attention decoder that models diverse temporal patterns from both multi-rank and multi-scale views. Experimental results on four popular benchmark datasets demonstrate that our proposed COOL provides state-of-the-art performance compared with the competitive baselines.

It is necessary to establish a spatio-temporal correlation model in the traffic data to predict the state of the transportation system. Existing research has focused on traditional graph neural networks, which use predefined graphs and have shared parameters. But intuitive predefined graphs introduce biases into prediction tasks and the fine-grained spatio-temporal information can not be obtained by the parameter sharing model. In this paper, we consider it is crucial to learn node-specific parameters and adaptive graphs with complete edge information. To show this, we design a model based on graph structure that decouples nodes and edges into two modules. Each module extracts temporal and spatial features simultaneously. The adaptive node optimization module is used to learn the specific parameter patterns of all nodes, and the adaptive edge optimization module aims to mine the interdependencies among different nodes. Then we propose a Decoupled Adaptive Graph Convolution Attention Network for Traffic Forecasting (DAGCAN), which relies on the above two modules to dynamically capture the fine-grained spatio-temporal relationships in traffic data. Experimental results on four public transportation datasets, demonstrate that our model can further improve the accuracy of traffic prediction.

No abstract available

No abstract available

Forecasting urban traffic states is crucial to transportation network monitoring and management, playing an important role in the decision-making process. Despite the substantial progress that has been made in developing accurate, efficient, and reliable algorithms for traffic forecasting, most existing approaches fail to handle sparsity, high-dimensionality, and nonstationarity in traffic time series and seldom consider the temporal dependence between traffic states. To address these issues, this work presents a Hankel temporal matrix factorization (HTMF) model using the Hankel matrix in the lower dimensional spaces under a matrix factorization framework. In particular, we consider an alternating minimization scheme to optimize the factor matrices in matrix factorization and the Hankel matrix in the lower dimensional spaces simultaneously. To perform traffic state forecasting, we introduce two efficient estimation processes on real-time incremental data, including an online imputation (i.e., reconstruct missing values) and an online forecasting (i.e., estimate future data points). Through extensive experiments on the real-world Uber movement speed data set in Seattle, Washington, we empirically demonstrate the superior forecasting performance of HTMF over several baseline models and highlight the advantages of HTMF for addressing sparsity, nonstationarity, and short time series. History: Accepted by Ram Ramesh, Area Editor for Data Science & Machine Learning. Funding: This research was supported by the Institute for Data Valorisation, the Interuniversity Research Centre on Enterprise Networks, Logistics and Transportation, the National Natural Science Foundation of China [Grants 12371456, 72101049, 72232001], the Sichuan Science and Technology Program [Grant 2024NSFJQ0038], and the Fundamental Research Funds for the Central Universities [Grant DUT23RC(3)045].

Due to the global trend towards urbanization, people increasingly move to and live in cities that then continue to grow. Traffic forecasting plays an important role in the intelligent transportation systems of cities as well as in spatio-temporal data mining. State-of-the-art forecasting is achieved by deep-learning approaches due to their ability to contend with complex spatio-temporal dynamics. However, existing methods assume the input is fixed-topology road networks and static traffic time series. These assumptions fail to align with urbanization, where time series are collected continuously and road networks evolve over time. In such settings, deep-learning models require frequent re-initialization and re-training, imposing high computational costs. To enable much more efficient training without jeopardizing model accuracy, we propose the Topological Evolution-aware Framework (TEAM) for traffic forecasting that incorporates convolution and attention. This combination of mechanisms enables better adaptation to newly collected time series while being able to maintain learned knowledge from old time series. TEAM features a continual learning module based on the Wasserstein metric that acts as a buffer that can identify the most stable and the most changing network nodes. Then, only data related to stable nodes is employed for re-training when consolidating a model. Further, only data of new nodes and their adjacent nodes as well as data pertaining to changing nodes are used to re-train the model. Empirical studies with two real-world traffic datasets offer evidence that TEAM is capable of much lower re-training costs than existing methods are, without jeopardizing forecasting accuracy.

Accurate traffic forecasting is more necessary than ever for transportation departments, especially given its significant role in traffic planning, management, and control. However, most existing methods struggle to address complex spatial correlations on road networks, nonlinear temporal dynamics, and difficult long‐term prediction. This article proposes a novel spatial temporal graph gated transformer (STGGT) to overcome these challenges. The suggested model differs from Google's transformer because it uses a hybrid architecture that integrates graph convolutional networks (GCNs), attention, and gated recurrent units (GRUs) instead of solely relying on attention. Specifically, STGGT uses GCNs to extract spatial dependencies, utilizes attention and GRUs to extract temporal dependencies, and handle long‐term prediction. Experiments indicate that STGGT outperforms the state‐of‐the‐art baseline models on two real‐world traffic datasets of 9%–40%. The proposed model offers a promising solution for accurate traffic forecasting, simultaneously addressing the challenges of complex spatial correlations, nonlinear temporal dynamics, and long‐term prediction.

Traffic forecasting is the foundation for intelligent transportation systems. Spatiotemporal graph neural networks have demonstrated state-of-the-art performance in traffic forecasting. However, these methods do not explicitly model some of the natural characteristics in traffic data, such as the multiscale structure that encompasses spatial and temporal variations at different levels of granularity or scale. To that end, we propose a Wavelet-Inspired Graph Convolutional Recurrent Network (WavGCRN) which combines multiscale analysis (MSA)-based method with Deep Learning (DL)-based method. In WavGCRN, the traffic data is decomposed into time-frequency components with Discrete Wavelet Transformation (DWT), constructing a multi-stream input structure; then Graph Convolutional Recurrent networks (GCRNs) are employed as encoders for each stream, extracting spatiotemporal features in different scales; and finally the learnable Inversed DWT and GCRN are combined as the decoder, fusing the information from all streams for traffic metrics reconstruction and prediction. Furthermore, road-network-informed graphs and data-driven graph learning are combined to accurately capture spatial correlation. The proposed method can offer well-defined interpretability, powerful learning capability, and competitive forecasting performance on real-world traffic data sets.

The Neural Controlled Differential Equation (NCDE) elegantly fuses dynamical systems with deep learning, unveiling profound potential for time series modeling. Harnessing the power of neural networks to sculpt the vector fields inherent to differential equations, this methodology introduces a seamless perspective for emulating spatial-temporal dynamical paradigms. In our research, the NCDE serves as the foundational architecture. Within this construct, we adeptly integrate both Temporal Convolutional Networks (TCNs) and Graph Neural Networks (GNNs) into a framework of continuous state representation to capture long-term spatial-temporal dependencies. This avant-garde modeling approach illuminates new avenues for nuanced modeling of spatial-temporal traffic dynamics, markedly augmenting the fidelity of traffic forecasting. Empirical evaluations conducted on three publicly accessible traffic flow datasets further underscore the superior efficacy of our proposed model in traffic forecasting.

Traffic flow forecasting is crucial for intelligent transportation systems. Currently, most models need to pay more attention to the delay (history) state to improve forecasting performance. In this paper, we propose a dynamic delay differential equation spatiotemporal network for traffic flow forecasting, named D3STN. First, our method integrates dynamic delay state optimization into delay differential equations to enhance delay state inputs and model forecasting performance. In addition, we propose a hybrid graph neural network and convolution multi-head attention mechanism. With the hybrid graph neural network, our method takes the multi-granularity correlation relationship into account to capture spatial characteristics from relevant nodes. With the convolution multi-head attention mechanism, our method balances the attention distribution between short-term and long-term attention. Empirical experiments are executed on highway traffic flow and metro flow datasets to evaluate the performance of our method. The results demonstrate that D3STN achieves significant advancements in traffic flow forecasting tasks. Compared with CorrSTN (baseline), D3STN makes improvements of 10.93%, 16.24% and 24.56% in terms of the MAE, RMSE and MAPE, respectively, on the HZME (outflow) dataset.

Traffic forecasting is a critical task in the field of Intelligent Transportation Systems. Previous research in traffic forecasting has primarily focused on integrating Graph Neural Networks with other models. However, as the networks utilized for handling spatio-temporal features become more complex, the improvements of these researches improve slowly. In this paper, we aim to capture the intricate spatio-temporal traffic patterns in a simple and efficient manner by introducing a framework called Spatio-Temporal Information Mixer (STIM). We analyze various information of spatio-temporal data and mix them into spatio-temporal embedding, incorporating time series information, periodicity information, and spatio-temporal position-aware information. Subsequently, we use two simple backbones, based on Multilayer Perceptron (MLP) or attention mechanisms, to validate the effectiveness of the spatio-temporal embedding. The proposed method in this paper achieves state-of-the-art performance in both accuracy and speed on the four authentic real-world traffic forecasting datasets.

No abstract available

Accurate Traffic Prediction is a challenging task in intelligent transportation due to the spatial-temporal aspects of road networks. The traffic of a road network can be affected by long-distance or long-term dependencies where existing methods fall short in modeling them. In this paper, we introduce a novel framework known as Spatial-Temporal Multi-Granularity Framework (STMGF) to enhance the capture of long-distance and long-term information of the road networks. STMGF makes full use of different granularity information of road networks and models the long-distance and long-term information by gathering information in a hierarchical interactive way. Further, it leverages the inherent periodicity in traffic sequences to refine prediction results by matching with recent traffic data. We conduct experiments on two real-world datasets, and the results demonstrate that STMGF outperforms all baseline models and achieves state-of-the-art performance.

Accurate traffic forecasting is crucial for the development of Intelligent Transportation Systems (ITS), playing a pivotal role in modern urban traffic management. Traditional forecasting methods, however, struggle with the irregular traffic time series resulting from adaptive traffic signal controls, presenting challenges in asynchronous spatial dependency, irregular temporal dependency, and predicting variable-length sequences. To this end, we propose an Asynchronous Spatio-tEmporal graph convolutional nEtwoRk (ASeer) tailored for irregular traffic time series forecasting. Specifically, we first propose an Asynchronous Graph Diffusion Network to capture the spatial dependency between asynchronously measured traffic states regulated by adaptive traffic signals. After that, to capture the temporal dependency within irregular traffic state sequences, a personalized time encoding is devised to embed the continuous time signals. Then, we propose a Transformable Time-aware Convolution Network, which adapts meta-filters for time-aware convolution on the sequences with inconsistent temporal flow. Additionally, a Semi-Autoregressive Prediction Network, comprising a state evolution unit and a semiautoregressive predictor, is designed to predict variable-length traffic sequences effectively and efficiently. Extensive experiments on a newly established benchmark demonstrate the superiority of ASeer compared with twelve competitive baselines across six metrics.

Traffic forecasting is essential in improving and maintaining safety and orderliness in intelligent transportation systems (ITS). As a deep learning approach, graph neural networks (GNN) based spatial-temporal association mining methods are promising in traffic forecasting. However, current GNN-based methods usually require a high number of training data, and when the sample volume is small, the performance of the model drops dramatically. The existing transfer methods can solve this problem by leveraging knowledge from other data-rich areas, but the domain adaption method with access to source data still faces the non-neglectable problem of private information leakage in the source area. A solution that can solve cross-area transfer without access to source data is still missing. In this paper, to fill the gap, we propose a Transferable Federated Inductive Spatial-Temporal Graph Neural Network (T-ISTGNN) framework to transfer spatial-temporal dependency information in cross-area data to accomplish traffic state forecasting. First, we introduce a multi-source model aggregation scheme based on federated learning to retain the traffic information of the source areas. Second, we propose a transfer method between source and target areas based on hypothesis transfer learning to achieve domain adaption under source domain data protection. Third, we propose a GNN-based method called Inductive Spatial-Temporal Graph Neural Network (ISTGNN) for traffic forecasting. Experiments on real-world datasets demonstrate that T-ISTGNN is capable of cross-area traffic state forecasting under the restriction of preserving the privacy of source areas.

Traffic flow prediction is a core task in intelligent transportation systems, but existing methods still face challenges in modeling complex spatiotemporal dependencies. This paper proposes an improved Graph Attention Network (GAT) that enhances prediction accuracy through a dual-path innovation mechanism. First, to address the insufficient feature extraction in the time dimension of traditional methods, we design a temporal-spatial separation architecture: the one-dimensional convolutional layer (Temporal Convolution) is introduced in the time dimension to capture the dynamic evolution patterns of traffic flow through local receptive fields. In the spatial dimension, we improve the graph attention mechanism by proposing a learnable dynamic edge weight matrix, breaking through the static limitations of traditional adjacency matrices. Second, we build a multi-scale feature fusion module that integrates spatial features at different time granularities through a multi-head attention mechanism. Experimental results show that on the PeMS dataset, the improved model reduces the Mean Square Error (MAE) and other metrics by over 50%. This study provides an interpretable improvement paradigm for spatiotemporal graph neural networks, and its lightweight nature makes it more suitable for practical deployment in transportation systems.

With the intensification of urban traffic congestion, how to improve the deployability of models while ensuring prediction accuracy has become a key challenge for intelligent transportation systems (ITS). This article proposes a new lightweight spatiotemporal prediction model - ST MLPWave. The model first uses Haar wavelet decomposition to decompose the traffic time series into low-frequency trends and high-frequency disturbances, and then models them separately through minimalist Temporal MLP, and achieves dynamic fusion through gating mechanism. In terms of spatial modeling, GraphConv Lite is introduced to capture the spatial correlation of road networks with only one adjacency propagation, avoiding the high complexity of traditional graph convolution structures. At the same time, this article designs a wavelet consistency loss function to constrain the prediction results from a frequency domain perspective, alleviate the problems of over smoothing and abnormal jitter, and further integrate exogenous features such as time periods and holidays to enhance the model's adaptability to periodic and sudden traffic patterns. The experimental results on synthetic datasets, METR-LA, PEMS-BAY and other real datasets show that ST-MLPWave is significantly better than existing mainstream methods in MAE, RMSE, MAPE and other indicators, and has the advantages of small parameter size, stable training, and easy reproducibility. This study provides a new approach for traffic prediction tasks that balances accuracy and lightweight, and has good engineering application prospect. Compared with existing hybrid models such as WaveNet or STGNN variants, ST-MLPWave emphasizes a frequency-domain constrained lightweight design, avoiding heavy temporal or attention-based structures while still maintaining competitive accuracy.

Accurate traffic prediction in large cities such as Los Angeles is increasingly necessary as cities expand and more vehicles are added to the roads. Using the METR-LA dataset. Using the METR-LA dataset, this study proposes a hybrid deep learning architecture that combines time and space modeling techniques to improve the accuracy and scalability of traffic flow predictions. The dataset consists of multivariate time series data from 207 loop detectors that record traffic speeds every five minutes with very high resolution. This study evaluates five potential model configurations: Long Short-Term Memory (LSTM), Transformer-based TSFormer, a combination of LSTM and TSFormer, Spatio-Temporal Graph Convolutional Network (STGCN), and a model combining STGCN and TSFormer. The evaluation conducted using three performance metrics Mean Absolute Error (MAE), Mean Squared Error (MSE), and Root Mean Squared Error (RMSE) were used to assess how well each model captures complex temporal and spatial relationships. Our results show that the LSTM+TSFormer hybrid model consistently outperforms all other models across all criteria. This model has the lowest MAE (0.0624) and RMSE (0.1204), meaning it is better at learning patterns that occur over time and patterns that occur rapidly. STGCN-based models are quite good at capturing spatial dependencies, but their performance improves when combined with attention-based TSFormer modules. The hybrid models introduced in this study overcome major limitations, including the narrow receptive range of recurrent networks and the inflexible spatial structures assumed in graph-based methods. This work offers important perspectives for developing forecasting models that are not only accurate and scalable but also transparent and adaptable. Future work may explore dynamic graph construction and multimodal input integration to further enhance adaptability in real-world applications.

Intelligent Transportation Systems (ITSs) have become pivotal in urban traffic management by utilizing traffic flow prediction, which aids in alleviating congestion and facilitating route planning. This study introduces the Linear Attention Based Spatial-Temporal Multi-Graph Convolutional Neural Network (LASTGCN), a novel deep learning model tailored for traffic flow prediction. LASTGCN incorporates a Multifactor Fusion Unit (MFF-unit) to dynamically integrate meteorological factors, an advanced multi-graph convolutional network for spatial correlations, and the Receptance Weighted Key Value (RWKV) block, which employs a linear attention mechanism for efficient processing of historical traffic data.The model achieves computational efficiency by using RWKV, which offers advantages over Transformer-based models in handling large-scale data while capturing complex dependencies. The model is designed to achieve computational efficiency, making it suitable for mid-term traffic management scenarios and potentially adaptable to real-time applications with further optimization. Experimental results using real-world highway traffic datasets indicate that LASTGCN outperforms several state-of-the-art methods in terms of accuracy and robustness, especially in long-term predictions. Additionally, integrating external factors such as weather conditions was found to significantly enhance the model’s predictive accuracy.

Exploring the network-level spatiotemporal traffic interplay through traffic pattern recognition or traffic prediction is essential for intelligent traffic systems. Prediction models commonly lack interpretability, while the static and unsupervised insights from pattern recognition underperform in dynamic and stochastic applications. Therefore, the authors developed a spatiotemporal factorized graph neural network (STF-GNN) framework to integrate traffic prediction and pattern recognition processes. Unlike prevailing prediction models, it shifts the prediction variables from traffic states to latent informative patterns. Specifically, graph-based and time-series models are applied to predict the spatial and temporal patterns, respectively, during the prediction horizons. Traffic predictions are derived based on the latent patterns, enhancing the model interpretability. In addition, decoupling the spatiotemporal correlations enables more stable and scalable pattern recognition. Thus, a pretrained pattern-driven model can be transferred to longer prediction horizons. The proposed framework was applied to two highway traffic datasets and two public transit datasets. Short-term prediction experiments reveal significant improvements in prediction and accuracy. Transfer learning experiments were also conducted. The predictors with superior accuracy in short-term scenarios perform effectively over extended horizons, demonstrating the proposed framework’s ability to provide precise predictions without compromising the validity of the recognized patterns. The patterns capture both predominant and marginal traffic characteristics during a relatively long horizon, supporting the prediction accuracy, interpretability, and scalability.

Traffic prediction, as a core technology of Intelligent Transportation Systems, plays a pivotal role in dynamic road network optimization and urban travel planning. However, the complex spatiotemporal characteristics of transportation networks pose significant challenges to precisely capturing their dynamic patterns. Existing methods predominantly rely on predefined static adjacency matrices and employ separate processing of spatial and temporal features, failing to adequately explore the intrinsic coupling relationships between them. To address these limitations, we propose an adaptive spatiotemporal dynamic graph convolutional network (AST-DGCN) for traffic prediction. Under the encoder-decoder architecture, the proposed model leverages node embedding techniques to extract high-dimensional features, generating time-evolving adaptive graphs through self-attention mechanisms. Concurrently, the model synergistically integrates dynamic graphs with gated recurrent units to achieve joint modeling of complex spatiotemporal dependencies. Furthermore, it introduces a dual-layer encoder-decoder residual correction module that effectively compensates for prediction errors, substantially enhancing forecasting accuracy. Experimental results on four public traffic datasets demonstrate that the AST-DGCN model achieves significant performance advantages over baseline methods across three critical evaluation metrics: root mean square error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE), thereby fully validating its superior predictive capabilities and competitive advantages.

No abstract available

No abstract available

No abstract available

Urban traffic flow management faces increasing challenges due to accelerating urbanization. Traffic data collected from roadside sensors contain complex temporal and spatial dependencies that interact simultaneously. Although Graph Neural Networks and Recurrent Neural Networks have been successful in capturing these dependencies, two critical issues remain: 1) Treating all traffic signals equally fails to capture the nuanced spatiotemporal dependencies hidden in time series data; 2) Dynamic traffic conditions hinder the accurate capture of local spatial dependencies, thereby limiting prediction accuracy and reliability. To address these challenges, we propose an innovative model, FDDSGCN. To resolve the first issue, we introduce Fractional Residual Decomposition, which effectively separates traffic data into spatial and temporal signals. For the second issue, we employ Dynamic Spatial-Temporal Graph Convolution with fractional-order weight adjustments to dynamically capture local dependencies. Additionally, the Long Short-Term Dependency module analyzes both long-term and short-term dependencies. Extensive experiments on three public datasets demonstrate the superior performance and practical value of our model.

Abstract Accurate traffic flow forecasting is critical for developing intelligent transportation systems. While Graph Neural Networks have achieved notable success in short-term forecasting, their performance significantly degrades over long-term horizons due to the complex entanglement of heterogeneous temporal patterns and the dynamic, non-local nature of spatial dependencies in road networks. To tackle these challenges, this study proposes a novel framework, termed Decoupled Spatiotemporal Graph Convolution with Probabilistic Sparse Self-Attention (DSGC-PSSA). In the temporal domain, a decoupling module is utilized to disentangle and highlight heterogeneous traffic flow patterns through frequency-domain analysis. Subsequently, a time-frequency fusion mechanism extracts complementary information from the decomposed features, thereby enhancing the model’s capability to capture complex temporal dynamics, particularly in long-term forecasting. To model the inherently dynamic nature of road networks, the spatial module of DSGC-PSSA adopts a hybrid graph learning strategy that integrates adaptive graph convolution and dynamic graph attention mechanisms. This design enables effective modelling of intricate spatial dependencies and supports flexible representations across both global and local scales. Additionally, the spatiotemporal convolution layer captures complex temporal dependencies through multi-scale modelling. In parallel, the probabilistic sparse self-attention mechanism facilitates dynamic integration of spatiotemporal features, thereby enabling more efficient long-term traffic flow forecasting. Extensive experiments conducted on five real-world traffic datasets demonstrate that DSGC-PSSA significantly outperforms existing state-of-the-art models, particularly in long-term forecasting.

Accurate traffic forecasting is essential for intelligent transportation system management and control. Due to the highly complex spatiotemporal (ST) correlation of real-world road networks, dynamic and long-term traffic prediction presents many challenges. We propose a traffic speed prediction model based on dynamic structural prior (DSP) ST graph attention networks. We provide a structural prior graph, namely, dual graph convolution, which combines spatial and contextual subgraphs to enable the discovery of the non-Euclidean spatial correlation and potential contextual similarity of road networks. Moreover, to dynamically extract the ST correlation, this article employs a multihead self-attention temporal convolution module to capture the temporal correlation and a graph attention convolution module to extract the spatial correlation. The prediction output is generated by stacking multiple ST blocks. Experimental results on two real-world traffic datasets demonstrate that DSP-ST outperforms existing mainstream baselines, which can provide references for traffic management departments.

ABSTRACT Existing traffic flow modeling approaches typically rely on real-time observations, such as road sensors or GPS trajectories, which constrain their research scope and application scenarios. This study proposes a novel cross-spatiotemporal graph-based network method to rapidly reconstruct traffic flow speed based on remote sensing images. The method is designed to address the challenges of traffic modeling in the absence of ground observation data. Combining high-resolution remote sensing imagery, vehicle object detection, and graph modeling technology, our approach could handle the discontinuous spatiotemporal graph information. The method incorporates two key modules: a two-layer masked structure mechanism and a cross-spatiotemporal attention computation. This innovative design enables the model to continuously synthesize learning from discontinuous remote sensing images and sparse ground-based sensor data during pre-training, optimizing its parameters and improving prediction accuracy over time. Once pre-trained, the graph model can directly estimate street-level traffic flow speed based solely on remote sensing images. Our results demonstrate state-of-the-art performance (MSE=40.117, MAE=4.768, RMSE=6.334, RSE=0.228), outperforming previous graph-based and sequence-based models. This study showcases the potential of utilizing remote sensing techniques to reconstruct traffic speed in urbanizing regions. It can even be used in scenarios lacking sufficient ground stations and with discontinuous remote sensing data, and enables low-cost, large-scale, and multi-temporal traffic flow speed reconstruction.

No abstract available

The spatiotemporal dynamics of traffic forecasting make it a challenging task. In recent years, by adapting to the topology of traffic networks where road segments serve as nodes, graph convolutional networks (GCNs) have been able to capture spatiotemporal dependencies, thereby improving traffic forecasting performance. However, there are two shortcomings of GCN-based methods: 1) existing methods treat the delays between nodes in the traffic network as equally important and fail to extract critical information effectively, leading to information redundancy, the introduction of irrelevant noise, and increased computational costs and 2) most methods overlook the issue that spatiotemporal correlations between nodes are inconsistent across different timescales. This article designs a new dynamic delay-aware multiscal spatiotemporal graph convolutional network (DDAMGCN) for traffic forecasting. Specifically, a dynamic delay-aware module is designed to identify key nodes and model the important delays from key nodes, so that the model focuses on key information and reduces computational cost. Additionally, a novel multiscale spatiotemporal graph convolution module is designed to achieve fine-grained modeling of the spatiotemporal correlation of different nodes at different timescales. Experiments on eight real traffic datasets verify the superiority of the proposed method compared to several state-of-the-art baselines.

As a typical spatiotemporal series prediction task, traffic flow prediction has found wide application in intelligent transportation systems (ITS). Despite some progress, several unresolved issues persist. Many existing works calculate the dependencies between nodes based on stable long-term traffic data. However, the short-term dependencies are dynamically changing over time, and neglecting them would cause a decrease in predictive performance. In this article, we propose a novel dynamic graph convolution and spatiotemporal self-attention (DGSTA) network for traffic flow prediction. Specifically, considering the large amount of short-term and the dynamic dependencies between nodes, we design a new dynamic graph convolution module, which generates adjacency matrices for each time step in a day to dynamically capture the changing short-term dependencies. Additionally, we utilize a multihead spatiotemporal self-attention module to, respectively, extract static spatial and temporal correlations between nodes. Furthermore, we design a sequential embedding to explicitly model the long-term correlation between nodes. Extensive experiments conducted on three real-world datasets demonstrate that DGSTA exhibits high competitiveness. The code and data are available at https://github.com/lzmmm30/DGSTA.

Accurate traffic flow prediction (TFP) is the basis for building intelligent transportation systems in smart cities. Existing methods usually focus on capturing the spatial dependencies of static graph structures, ignoring the explicit road network structure and failing to mine the embedded spatiotemporal dependencies and characteristics of traffic network structures. To this end, we propose an attention-based spatiotemporal adaptive graph diffusion convolutional network (AST-AGDCN) to accurately describe the spatial structure and fully characterize the spatiotemporal dependencies of the traffic data for TFP. Specifically, we designed a spatiotemporal self-attention mechanism module for extracting the potential effects of temporal patterns and spatial correlations among traffic network nodes and mining the degree of influence of the changing spatiotemporal location. A road network connectivity graph modeling module was then constructed to adaptively extract spatial relationships between nodes to learn external information. Next, a diffusion convolution network was built to model the directionality and dynamics of the spatial road network structure, and the complex spatial structural properties between the regions were aggregated into a feature matrix to discover hidden spatial pattern correlations effectively. Finally, temporal features of the dynamic traffic flow were explicitly modeled using a bidirectional gated recurrent unit, and weights were fused to the results of the recent, daily, and weekly periodic segments for TFP. The extensive experimental results demonstrated that AST-AGDCN was superior to comparable models in prediction accuracy on real-world datasets.

Traffic forecasting is considered a cornerstone of smart city development. A key challenge is capturing the long-term spatiotemporal dependencies of traffic data while improving the model’s generalization ability. To address these issues, various sophisticated modules are embedded into different models. However, this approach increases the computational cost of the model. Additionally, adding or replacing datasets in a trained model requires retraining, which decreases prediction accuracy and increases time cost. To address the challenges faced by existing models in handling long-term spatiotemporal dependencies and high computational costs, this study proposes an enhanced pre-training method called the Improved Spatiotemporal Diffusion Graph (ImPreSTDG). While existing traffic prediction models, particularly those based on Graph Convolutional Networks (GCNs) and deep learning, are effective at capturing short-term spatiotemporal dependencies, they often experience accuracy degradation and increased computational demands when dealing with long-term dependencies. To overcome these limitations, we introduce a Denoised Diffusion Probability Model (DDPM) as part of the pre-training process, which enhances the model’s ability to learn from long-term spatiotemporal data while significantly reducing computational costs. During the pre-training phase, ImPreSTDG employs a data masking and recovery strategy, with DDPM facilitating the reconstruction of masked data segments, thereby enabling the model to capture long-term dependencies in the traffic data. Additionally, we propose the Mamba module, which leverages the Selective State Space Model (SSM) to effectively capture long-term multivariate spatiotemporal correlations. This module enables more efficient processing of long sequences, extracting essential patterns while minimizing computational resource consumption. By improving computational efficiency, the Mamba module addresses the challenge of modeling long-term dependencies without compromising accuracy in capturing extended spatiotemporal trends. In the fine-tuning phase, the decoder is replaced with a forecasting header, and the pre-trained parameters are frozen. The forecasting header includes a meta-learning fusion module and a spatiotemporal convolutional layer, which facilitates the integration of both long-term and short-term traffic data for accurate forecasting. The model is then trained and adapted to the specific forecasting task. Experiments conducted on three real-world traffic datasets demonstrate that the proposed pre-training method significantly enhances the model’s ability to handle long-term dependencies, missing data, and high computational costs, providing a more efficient solution for traffic prediction.

The graph-based traffic forecasting is generally realized on the assumption of sufficient data, which could be impractical in the regions without well-deployed mobile sensors or data-processing facilities. Recent studies have developed a solution with the cross-region transfer learning, i.e. transferring traffic knowledge from the source regions to target ones, whose traffic data and computing resources are limited. Nevertheless, relevant issues, including initialization selection and domain adaptation, have not been effectively tackled in the cross-region graph-based traffic forecasting. This paper proposes the cluster-granularity spatiotemporal transfer (CGSTT), which transfers the cluster-granularity knowledge from the source region to target one for the cross-region graph-based traffic forecasting, as not all source knowledge is positive to the target region. Additionally, the domain adaptation is achieved by the dual alignment consisting of the covariate alignment and label alignment of the source/target data, making the proposed CGSTT adapt to the target region efficiently. The superiority of the proposed method over ten compared baseline methods for both short-term and long-term predictions is demonstrated by the conducted experiments on four tasks, which show that it outperforms the state-of-the-art method by achieving an 8.89% average improvement in forecasting accuracy. The PyTorch implementation of the CGSTT is available at https://github.com/canyangguo/CGSTT.

Traffic prediction is a cornerstone of intelligent transportation systems (ITSs). The effectiveness of existing spatiotemporal graph neural networks (STGNNs) heavily relies on the independent identically distributed (i.i.d.) assumption of traffic data, which is frequently violated in practice because of distribution shifts owing to exogenous factors. While learning features that remain stable across all environments is promising for modeling robust frameworks, the fundamental challenge involves the decomposition of invariant features from the dynamic nature of spatiotemporal dependencies. In this article, we propose the disentangled spatiotemporal (DIST) graph neural networks, a novel framework for robust traffic forecasting considering distribution shifts. In DIST, latent invariant variables are explicitly decoupled from dynamically evolving spatiotemporal dependencies, enabling the learning of topology-agnostic representations resilient to distribution shifts. Specifically, we formulate a causality-driven learning objective that guides the separation of invariant variables from various exogenous factors. We then propose a spatiotemporal graph modeling module that can adaptively capture spatiotemporal dependencies in evolving traffic systems. Furthermore, we present a graph perturbation module to simulate topology variations during training, thereby encouraging the model to identify perturbation-sensitive dependencies and infer invariant and variant features for prediction and intervention tasks. The prediction risk and its variance on multiple interventional distributions are minimized in our learning strategy, allowing the model to identify invariant features, thus improving its robustness. The results of comprehensive real-world experiments demonstrate the superiority of our approach. The source code is available: https://github.com/tingwang25/DIST.

Traffic prediction is vital to traffic planning, control, and optimization, which is necessary for intelligent traffic management. Existing methods mostly capture spatiotemporal correlations on a fine-grained traffic graph, which cannot make full use of cluster information in coarse-grained traffic graph. However, the flow variation of clusters in the coarse-grained traffic graph is more stable compared with nodes in the fine-grained traffic graph. And the flow variation of a fine-grained node is generally consistent with the trend of the cluster to which the node belongs. Thus information in the coarse-grained traffic graph can guide feature learning in the fine-grained traffic graph. To this end, we propose a Spatiotemporal Multiscale Graph Convolutional Network (SMGCN) that explores spatiotemporal correlations on a multiscale graph. Specifically, given a fine-grained traffic graph, we first generate a coarse-grained traffic graph by graph clustering, and extract spatiotemporal correlations on both fine-grained and coarse-grained traffic graphs. Then we propose a cross-scale fusion (CF) to implement information diffusion between the fine-grained and coarse-grained traffic graphs. Moreover, we employ an adaptive dynamic graph convolution network to mine both static and dynamic spatial features. We evaluate SMGCN on real-world datasets and obtain a $1.18\% -3.32\%$ improvement over state-of-the-arts.

Spatiotemporal traffic flow prediction is a fundamental task in intelligent transportation systems and is crucial for promoting efficient and sustainable urban mobility, especially under increasingly complex and rapidly evolving traffic conditions. To overcome the challenges of modeling high-order spatial dependencies and heterogeneous temporal patterns, this study develops a novel Hierarchical Spatiotemporal Graph–Hypergraph Network (HSTGHN). For spatial representation learning, a hypergraph neural module is employed to capture high-order interactions across the road network, while a hypernode mechanism is designed to characterize complex correlations among multiple road segments. Furthermore, an adaptive adjacency matrix is constructed in a data-driven manner and enriched with prior knowledge of bidirectional traffic flows, thereby enhancing the robustness and accuracy of graph structural representations. For temporal modeling, HSTGHN integrates the complementary strengths of Gated Recurrent Units (GRUs) and Transformers: GRUs effectively capture local sequential dependencies, whereas Transformers excel at modeling global dynamic patterns. This joint mechanism enables comprehensive learning of both short-term and long-term temporal dependencies. Extensive experiments on multiple benchmark datasets demonstrate that HSTGHN consistently outperforms state-of-the-art baselines in terms of prediction accuracy and stability, with particularly significant improvements in long-term forecasting and highly dynamic traffic scenarios. These improvements provide more reliable decision support for intelligent transportation systems, contributing to enhanced traffic efficiency, reduced congestion, and ultimately more sustainable urban mobility.

Traffic flow prediction is a prominent research area in intelligent transportation systems, significantly contributing to urban traffic management and control. Existing methods or models for traffic flow prediction predominantly rely on a fixed-graph structure to capture spatial correlations within a road network. However, the fixed-graph structure can restrict the representation of spatial information due to varying conditions such as time and road changes. Drawing inspiration from the attention mechanism, a new prediction model based on the mixed-graph neural network is proposed to dynamically capture the spatial traffic flow correlations. This model uses graph convolution and attention networks to adapt to complex and changeable traffic and other conditions by learning the static and dynamic spatial traffic flow characteristics, respectively. Then, their outputs are fused by the gating mechanism to learn the spatial traffic flow correlations. The Transformer encoder layer is subsequently employed to model the learned spatial characteristics and capture the temporal traffic flow correlations. Evaluated on five real traffic flow datasets, the proposed model outperforms the state-of-the-art models in prediction accuracy. Furthermore, ablation experiments demonstrate the strong performance of the proposed model in long-term traffic flow prediction.

The morphological diversity, referring to the variations in traffic network topologies defined in this paper, often emerges and brings difficulties in successfully transferring a pre-trained prediction model from one traffic network to another. Moreover, most existing research primarily assumes that traffic data in source and target networks follow independent and identically distributed (i.i.d.) patterns, which is usually not consistent with real-world situations, particularly when considering morphological diversity. For this inconsistency, many efforts have been made, but they mainly concentrate on temporal aspects, which significantly differ from traffic prediction due to spatial and temporal correlations among road segments, influenced by variations in road topology and traffic behavior. This paper introduces a causality-based spatiotemporal out-of-distribution (OOD) generalization method, which is adaptable to most GNNs for diverse, large-scale, dynamic traffic systems with zero-shot. Furthermore, to enhance the generalization and adaptability of the proposed method, we introduce graph matching and equal-sized graph partitioning to alleviate spatial shift between the source and target traffic networks, reduce and align the scale of the networks. Experiments carried out on traffic flow datasets demonstrate that our method significantly improves the performance of various GNN-based traffic predictors in the situation of morphological diversity, achieving a maximum reduction in MAE of 33.08%. Compared to other OOD-driven baselines, our approach also shows a notable improvement, with up to a 40.58% decrease in MAE.

Accurate traffic speed prediction plays a key role in transportation planning. Inspired by advanced natural language processing (NLP) techniques in speech recognition, this paper proposes a bidirectional spatio-temporal gated graph convolutional network (BiSTGG) based on the attention mechanism for urban traffic speed prediction. The model analogizes the average daily traffic speed with word frequencies in speech, thus establishing a novel semantic perspective for traffic modeling. In order to comprehensively assess the similarity between traffic speed trajectories, the model introduces a dynamic temporal regularization (DTW) algorithm, which enables the construction of dynamic semantic similarity subgraphs that go beyond the traditional spatial proximity modeling approach. Meanwhile, the model applies the multi-head graph attention mechanism to ensure that the complex and diverse interactions in the road network can be captured effectively. In addition, the model introduces spatial identification vectors to enhance robustness and direct attention to focus on key features. The model also incorporates a bi-directional gated recurrent unit (Bi-GRU) structure with an attention mechanism to capture the complex forward and backward time dependencies in traffic patterns. Experiments on two large-scale real datasets show that BiSTGG performs well in terms of prediction accuracy and robustness, and significantly improves real-time traffic speed prediction.

Traffic flow forecasting is a fundamental task in intelligent transportation systems, directly supporting traffic control, congestion mitigation, and urban mobility planning. However, prediction remains difficult owing to nonlinear dynamics, rapidly changing spatiotemporal dependencies, and the integration of heterogeneous data. This paper proposes ASISTGCRN, an attention-based spatiotemporal graph convolutional recurrent network that introduces a tri-cycle segmentation strategy to capture recent, daily, and weekly periodic patterns. Central to the framework is a spatiotemporal block that integrates multi-head temporal and spatial attention with Dynamic Time Warping to account for both dynamic local dependencies and remote functional similarities. Adaptive graph convolution and gated recurrent units are employed to jointly model spatial structures and temporal sequences, while Transformer- and Informer-based attention layers are further applied to capture long-range dependencies, yielding two model variants, T-ASISTGCRN and I-ASISTGCRN. Extensive experiments on four widely used benchmark datasets (PEMS03, PEMS04, PEMS07M, PEMS08) demonstrate that ASISTGCRN consistently outperforms 17 baseline approaches across MAE, RMSE, and MAPE metrics. Ablation studies further verify the contribution of the tri-cycle segmentation, spatiotemporal block, and attention modules. These results indicate that the proposed framework offers improved robustness and accuracy in traffic flow prediction and has practical relevance for the design of advanced traffic management and planning strategies in complex urban environments.

Traffic flow prediction is the key to accurate urban traffic control and the basis for developing intelligent transportation systems. Recent studies have made substantial progress in traffic prediction by modelling complex spatiotemporal graph topology and considering sensors as road network nodes. However, the current spatiotemporal graph neural network model is limited by its structure. It can only utilize short-range traffic flow data and cannot effectively extract the long-term trend of complex traffic flow and periodic features in traffic patterns. To address the above problems, we propose a Transformer-based long-term traffic flow prediction framework, “Transformer-based spatiotemporal graph attention network”. First, the model utilizes the Transformer coding layer to learn compressed and context-rich subsequence temporal representations from long-term sequences. Then, the model designs a multi-scale gated temporal convolution module to identify and extract long-term trend features of traffic flow from the subsequence time representations. Next, the model constructs a multi-granularity random graph attention module to capture the periodic features of traffic flow from the subsequence time representations and extracts the short-term trend features present in the long-time series using the STGNN model. Finally, the model fuses the extracted long-term trends, periodic features and short-term trends to obtain the final prediction results. Experimental results on two real-world traffic flow datasets show that the model outperforms the baseline model and makes accurate long-term predictions.

No abstract available

Amidst rapid urbanization, traffic congestion and frequent accidents pose significant challenges to urban development. In this scenario, precise traffic prediction is imperative for enhancing the efficiency and safety of traffic management systems. Traditional approaches often fail to capture the complex spatiotemporal correlations and multi-level temporal dynamics of traffic data, rendering them inadequate for long-term traffic prediction. To address these limitations, this study introduces a novel approach employing an interpretable multi-scale adaptive dynamic spatiotemporal graph convolutional network. This method dynamically extracts latent relationships through a graph learning module to enhance traffic prediction accuracy and augments interpretability through a thorough analysis of data flow dynamics. Moreover, the model incorporates a dynamic graph construction technique within the prediction module to better model complex traffic scenarios, and introduces a scale fusion method to adaptively select different hierarchical levels of analysis. Comprehensive experiments on four real-world datasets-England, METR-LA, PEMSD4, and PEMSD8-demonstrate that our proposed model surpasses the best baselines by 32.6%, 7.37%, 25.9%, and 0.7% in mean absolute error, respectively, achieving an average improvement of 16.65%. Additionally, by analyzing the graph convolutional network node weight parameters and monitoring weight variations across different layers, we elucidate the model’s decision-making process, thereby providing an interpretative analysis of its functionality.

Sensors are always sparsely distributed in traffic networks due to high deployment costs, sensor damage, etc. Insufficient data may affect our perception of traffic scenarios, resulting in the inability of intelligent transportation systems (ITS) to efficiently perform traffic monitoring and scenario decisions. Spatial interpolation methods are used to infer the status of locations where no sensors are deployed. However, existing methods still have the following limitations: (1) Mainly interpolated unobserved nodes by extracting spatiotemporal dependencies between nodes, but ignored complex interaction patterns in feature dimensions. (2) Recent methods generally used standard TCN to extract temporal correlations, which is affected by abnormal data. (3) When extracting spatial correlations, the use of deep GCN layer can lead to over-smoothing problem. To mitigate these limitations, we propose a novel spatial interpolation model, namely, Spatiotemporal Trend Fusion Feature Graph Convolution Network (STFGCN). Specifically, a novel feature graph convolution network is used to capture complex interaction patterns. Secondly, the trend capture branch is used to alleviate the impact of abnormal data on TCN. Finally, a dual-stage spatial module is used to solve the degradation of detailed feature representations in deep GCN layer. Experimental results on six real traffic datasets demonstrate that our method outperforms state-of-the-art baseline models.

Traffic volume estimation is a fundamental task in Intelligent Transportation Systems (ITS). The highly unbalanced and asymmetric spatiotemporal distribution of traffic flow combined with the sparse and uneven deployment of sensors pose significant challenges for accurate estimation. To address these issues, this paper proposes a novel traffic volume estimation framework. It combines a dynamic adjacency matrix Graph Convolutional Network (GCN) with a multi-scale transformer structure to capture spatiotemporal correlation. First, an adaptive speed-flow correlation module captures global road correlations based on historical speed patterns. Second, a dynamic recurrent graph convolution network is used to capture both short- and long-range correlations between roads. Third, a multi-scale transformer module models the short-term fluctuations and long-term trends of traffic volume at multiple scales, capturing temporal correlations. Finally, the output layer fuses spatiotemporal correlations to estimate the global road traffic volume at the current time. Experiments on the PEMS-BAY dataset in California show that the proposed model outperforms the baseline models and achieves good estimation results with only 30% sensor coverage. Ablation and hyperparameter experiments validate the effectiveness of each component of the model.

Short-term traffic speed prediction under limited historical data is crucial for intelligent transportation systems, enabling real-time traffic management and congestion mitigation. However, existing methods relying on long lookback window to forecast struggle to capture scenarios where the current traffic state depends more on recent patterns than on distant historical data. To address this challenge, we propose the Spatiotemporal Visibility Graph Network (STVGN), a novel framework that combines visibility graph theory with Graph Neural Networks (GNN). The core of STVGN is the STVGN-Embedding layer, which is designed to enhance short-term spatiotemporal feature representation. This layer leverages the visibility graph’s ability to capture transient patterns and integrates GNNs to model intrinsic relationships within the visibility graph derived from time series data. In Combination with a transformer architecture, STVGN-Embedding extracts complex spatiotemporal features, while the transformer uncovers inherent relationships between temporal and spatial dimensions. To the best of our knowledge, this is the first study to introduce visibility graph theory as an embedding layer within a transformer framework. Experiments on three benchmark datasets, covering urban and freeway traffic scenarios, demonstrate STVGN’s effectiveness, achieving state-of-the-art performance with improvements of 4.8%–16.8% in RMSE and 8.4%–13.3% in MAE over existing SOTA methods. These results highlight the potential of visibility graph-based embeddings to address challenges posed by limited historical data and to capture intricate traffic patterns.

No abstract available

No abstract available

No abstract available

The characteristics of multivariate heterogeneity in traffic flow forecasting exhibit significant variation, heavily influenced by spatio-temporal dynamics and unforeseen events. To address this challenge, we propose a spatio-temporal fusion graph neural network based on dynamic sparse graph convolution GRU for traffic flow forecast (STFDSGCN), which incorporates a spatio-temporal attention fusion scheme with a gating mechanism. The dynamic sparse graph convolution gated recurrent unit (DSGCN-GRU) in this model is a novel component that integrates adaptive dynamic sparse graph convolution into the gated recurrent network to simulate the diffusion of information within a dynamic spatial structure. This approach effectively captures the heterogeneous and local features of spatial data, further reflecting the irregularities and dynamic variability inherent in spatial information. By leveraging spatio-temporal attention through the gating mechanism, the model enhances its understanding of both local and global spatio-temporal characteristics. This enables a unified representation of multi-scale and long-range spatio-temporal patterns and strengthens the model’s ability to respond to long-term traffic flow forecasting and traffic emergencies. Extensive experiments on two real-world datasets demonstrate that, compared to advanced methods that lack sufficient multivariate heterogeneous feature extraction and do not account for traffic emergencies, the STFDSGCN model improves the average absolute error (MAE), root mean square error (RMSE), and average absolute percentage error (MAPE) by 4.01%, 1.33%, and 1.03%, respectively, achieving superior performance.

The Urban traffic flow is affected by both internal supply and demand changes and external random disturbances, and during its continuous spatiotemporal propagation, these factors overlap with each other, presenting a highly non-linear and complex spatiotemporal pattern, which poses a huge challenge to traffic flow prediction. In response to the above challenges, this paper proposes a novel Spatio-Temporal Graph neural network with Multi-timeScale (abbreviated as STGMS). In STGMS, a multi-timescale feature decomposition strategy was designed to decompose the traffic flow into signals at multiple timescales and residuals. A unified spatio-temporal feature encoding module was designed to integrate the spatiotemporal features of traffic flow and the interaction features of multi-timescale traffic flows. Finally, the mapping from the multi-timescale spatiotemporal feature encoding to the future traffic flow was learned. We conducted numerous experiments on four real-world datasets and compared them with eleven baseline models from the past three years. The results show that the performance of our model outperforms the current state-of-the-art baseline models. On the four datasets, the average improvement rates of the three prediction accuracy metrics, namely the Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Mean Absolute Percentage Error (MAPE), reach 17.69%, 15.65%, and 10.30% respectively.

Predicting traffic is the main duty of an intelligent transportation system (ITS). Precise traffic forecasts can significantly enhance the use of public funds. However, the dynamic and complex nature of spatio-temporal relationships presents significant challenges. Most current methods utilize static adjacency matrices, leading to reduced forecasting accuracy and precision. This approach fails to account for the complex spatio-temporal correlations that interact simultaneously. In order to show how different spatio-temporal correlations change over time in the traffic flow network, this study suggests a unified simultaneous Multi Fusion Graph Network (DMFGNet) model. The goal of the suggested DMFGNet model is to identify dynamic spatio-temporal linkages between various regions. Meanwhile, we propose a model, the Spatio-Temporal Attention Unit (STAU), to control the weights of neighbor aggregation. It is capable of meticulously combining spatio-temporal characteristics from different neighbors. We tested the model for both real-time and pre-processed predictions using a combination of edge and cloud infrastructure. This setup performs prediction tasks at the edge layer and conducts training in the remote cloud. This approach guarantees the use of only relevant data for model training and prediction-making, thereby boosting the system’s overall effectiveness. This approach not only optimizes resource allocation but also aids in reducing latency and enhancing the overall performance of cloud-based prediction models, potentially enhancing the capabilities of consumer technology and electronics solutions. We carefully tested and evaluated two large real-world traffic flow datasets to show that the proposed method works and is useful. The results of the tests show that the suggested model is better than the current best baseline methods. Additionally, the results demonstrate the effectiveness and usefulness of the recommended strategy.

Accurate traffic prediction is crucial for urban traffic management. Spatial-temporal graph neural networks, which combine graph neural networks with time series processing, have been extensively employed in traffic prediction. However, traditional graph neural networks only capture pairwise spatial relationships between road network nodes, neglecting high-order interactions among multiple nodes. Meanwhile, most work for extracting temporal dependencies suffers from implicit modeling and overlooks the internal and external dependencies of time series. To address these challenges, we propose a Geometric Algebraic Multi-order Graph Neural Network (GA-MGNN). Specifically, in the temporal dimension, we design a convolution kernel based on the rotation matrix of geometric algebra, which not only learns internal dependencies between different time steps in time series but also external dependencies between time series and convolution kernels. In the spatial dimension, we construct a tokenized hypergraph and integrate dynamic graph convolution with attention hypergraph convolution to comprehensively capture multi-order spatial dependencies. Additionally, we design a segmented loss function based on traffic periodic information to further improve prediction accuracy. Extensive experiments on seven real-world datasets demonstrate that GA-MGNN outperforms state-of-the-art baselines.

No abstract available

No abstract available

Accurate prediction of traffic flow plays an important role in maintaining traffic order and traffic safety, which is a key task in the application of intelligent transportation systems (ITS). However, the urban road network has complex dynamic spatial correlation and nonlinear temporal correlation, and achieving accurate traffic flow prediction is a highly challenging task. Traditional methods use sensors deployed on roads to construct the spatial structure of the road network and capture spatial information by graph convolution. However, they ignore that the spatial correlation between nodes is dynamically changing, and using a fixed adjacency matrix cannot reflect the real road spatial structure. To overcome these limitations, this paper proposes a new spatial-temporal deep learning model: gated fusion adaptive graph neural network (GFAGNN). GFAGNN first extracts long-term dependencies on raw data through stacking expansion causal convolution, Then the spatial features of the dynamics are learned by adaptive graph attention network and adaptive graph convolutional network respectively, Finally the fused information is passed through a lightweight channel attention to extract temporal features. The experimental results on two public data sets show that our model can effectively capture the spatiotemporal correlation in traffic flow prediction. Compared with GWNET-conv model on METR-LA dataset, the three indexes in the 60-minute task prediction improved by 2.27%,2.06% and 2.13%, respectively.

Recently, mobile crowdsensing (MCS) has emerged as a promising solution for traffic state estimation (TSE), which provides real-time and accurate traffic information for supporting diversified intelligent transportation systems (ITSs) applications. However, the prohibitive overhead of collecting massive data in vehicular networks limits the available data amount, while the sparsification of MCS data incurs instability and degrades TSE accuracy. To this end, this article proposes a novel sparse MCS framework to facilitate cost-effective TSE, which utilizes a small number of vehicular MCS participants distributed across all regions as data sources. By utilizing spatial and temporal correlations of traffic flow, an innovative spatiotemporal deep learning model, namely, transformer graph attentional sample and aggregate (TGASA) neural network, is proposed to improve the TSE accuracy with sparse MCS data. Specifically, we design an incorporated graph neural network (GNN) to aggregate the spatial correlation by taking both node features and edge properties into account. And, the transformer neural network architecture is applied to capture the temporal correlation. Extensive simulation results based on real-world data sets demonstrate that the proposed framework can significantly address the instability incurred by the sparsification of MCS data and effectively achieve a more accurate TSE.

Traffic flow prediction is a non-negligible part of intelligent transportation and mobility. Unfortunately, the unique non-linearity and complex spatial-ST-correlation of transport flow data suggest considerable challenges in prediction. The dynamic interaction of multiple spatial relations greatly influences traffic flow prediction. However, the existing spatial-temporal prediction algorithms are based on graph convolution to capture global or heterogeneous relationships, and simpler graph convolution models cannot accurately capture complex dynamic spatial relationships. To address the issues as mentioned above, this study proposes an attention-based multi-graph dynamic spatial-temporal prediction model ADMSTGCN to capture a variety of dynamic interaction relationships in traffic flow. First, we use a distance graph to explore the relationships between adjacent distances and use a semantic graph to mine spatial relationships between nodes that are far apart but have similar relationships, then fuse these two graphs to obtain a fusion graph with multiple spatial interaction relationships. The correlations between different neighbors are then further learned through a dynamic multi-graph spatial-temporal learning module that aggregates the features of different neighbors through gated graph convolution and attention mechanisms to capture various dynamic and complex spatial-temporal interactions. Experimental evaluations show that the framework proposed outperforms existing methods with better results in the analysis performed with publicly available datasets and also demonstrates the importance of capturing multiple interactions of spatial-temporal relationships.

Traffic flow analysis, prediction and management are keystones for building smart cities in the new era. With the help of deep neural networks and big traffic data, we can better understand the latent patterns hidden in the complex transportation networks. The dynamic of the traffic flow on one road not only depends on the sequential patterns in the temporal dimension but also relies on other roads in the spatial dimension. Although there are existing works on predicting the future traffic flow, the majority of them have certain limitations on modeling spatial and temporal dependencies. In this paper, we propose a novel spatial temporal graph neural network for traffic flow prediction, which can comprehensively capture spatial and temporal patterns. In particular, the framework offers a learnable positional attention mechanism to effectively aggregate information from adjacent roads. Meanwhile, it provides a sequential component to model the traffic flow dynamics which can exploit both local and global temporal dependencies. Experimental results on various real traffic datasets demonstrate the effectiveness of the proposed framework.

No abstract available

No abstract available

We all depend on mobility, and vehicular transportation affects the daily lives of most of us. Thus, the ability to forecast the state of traffic in a road network is an important functionality and a challenging task. Traffic data is often obtained from sensors deployed in a road network. Recent proposals on spatial-temporal graph neural networks have achieved great progress at modeling complex spatial-temporal correlations in traffic data, by modeling traffic data as a diffusion process. However, intuitively, traffic data encompasses two different kinds of hidden time series signals, namely the diffusion signals and inherent signals. Unfortunately, nearly all previous works coarsely consider traffic signals entirely as the outcome of the diffusion, while neglecting the inherent signals, which impacts model performance negatively. To improve modeling performance, we propose a novel Decoupled Spatial-Temporal Framework (DSTF) that separates the diffusion and inherent traffic information in a data-driven manner, which encompasses a unique estimation gate and a residual decomposition mechanism. The separated signals can be handled subsequently by the diffusion and inherent modules separately. Further, we propose an instantiation of DSTF, Decoupled Dynamic Spatial-Temporal Graph Neural Network (D 2 STGNN), that captures spatial-temporal correlations and also features a dynamic graph learning module that targets the learning of the dynamic characteristics of traffic networks. Extensive experiments with four real-world traffic datasets demonstrate that the framework is capable of advancing the state-of-the-art.

With the advancement of modern UAV technology, UAVs have become integral to creating traffic management monitoring systems. Additionally, UAV-based traffic monitoring systems can predict traffic flow by integrating machine learning methods. Specifically, traffic flow data contains both spatial and temporal information, which can be effectively processed by graph neural networks (GNNs). However, GNNs often face the challenge of oversmoothing, which hinders their ability to capture complex structures in the data. The Spatio-Temporal Graph Ordinary Differential Equations (STGODE) model addresses this issue by introducing Neural Ordinary Differential Equations (NODEs) to construct deeper GNNs. Despite this, STGODE relies on initially predefined semantic neighborhood matrices, which do not adapt well to the dynamic nature of traffic information. To overcome this limitation, we propose an evolutionary graph neural network for traffic prediction, capable of continuously updating the semantic adjacency matrix throughout the training process. This dynamic evolution of the semantic adjacency matrix allows it to adapt to the features and semantic relations of the current data, enhancing its ability to capture the complexity and variability of traffic patterns. We validate our approach through experiments on several real-world datasets, demonstrating that our method outperforms state-of-the-art benchmarks.

Spatial-temporal network traffic prediction remains a formidable challenge due to the intricate spatial relationships and dynamic temporal patterns characteristic of individual nodes. Traditional regression methods fall short in handling such graph-structured data effectively. In recent developments, Graph Neural Networks (GNNs) have shown promise in modeling these complex spatial-temporal interactions. Despite their potential, existing GNN-based approaches exhibit notable limitations: (1) they typically rely on fixed spatial adjacency matrices, which overlook fuzzy latent temporal dependencies; and (2) they often process spatial and temporal information independently, resulting in a loss of joint fuzzy dependencies or restricting the model to either global or local patterns alone. To overcome these issues, we introduce the Fuzzy Spatial-Temporal Fusion Graph Neural Network (FSTFGNN) based on Fuzzy Rough Sets. Our approach constructs a dynamic, data-driven fuzzy spatial-temporal fusion matrix to capture underlying low-level spatial-temporal joint relationships. Additionally, a fuzzy global-local unified GNN layer is incorporated to simultaneously learn global-local spatialtemporal dependencies. Experimental results on two real-world datasets validate the efficacy of the proposed FSTFGNN framework.

No abstract available

No abstract available

No abstract available

No abstract available

INTRODUCTION: Adaptive Traffic Signal Optimisation (ATSO) is a challenging problem for urban traffic networks, having important implications for congestion reduction, traffic efficiency, and environmental conservation. Conventional traffic signal control techniques, i.e., fixed-time and rule-based control, fail to respond to dynamic traffic behaviour efficiently. OBJECTIVES: Recent developments in Reinforcement Learning (RL) have been promising for ATSO but are plagued by poor scalability, lack of coordination in multi-intersection networks, and inefficiency in dealing with continuous action spaces. METHODS: Furthermore, most RL-based solutions are based on simplistic state representation and fail to incorporate complex interdependencies between traffic signals. Considering these limitations, this paper introduces a new framework, Multi-Agent Soft Actor-Critic with Graph Attention Networks (MASAC-GAT), which unites the sample efficiency and stability of Soft Actor-Critic (SAC) with the relational modelling ability of Graph Attention Networks (GATs). RESULTS: The proposed method exhibited significant performance gains on three important traffic metrics: Signal Adjustment Efficiency (92%), Average Waiting Time (20–35 seconds), and Congestion Prediction Accuracy (93%), outperforming DQL, PPO, A2C, GNN-based variants, and knowledge sharing DDPG (KS-DDPG). Through minimised redundant signal changes and reduced vehicle delays, the method ushers in the next generation of smart transportation systems. CONCLUSION: The proposed method facilitates decentralised yet coordinated control of traffic signals by utilising local observations and global context. The proposed method unites real-time traffic observations, e.g., traffic volume, vehicle speeds, weather, accident reports, and signal status, into a customised OpenAI Gym environment for training and evaluation.  

Temporal traffic prediction is critical for ITS yet remains challenging in handling complex spatio-temporal dynamics of traffic systems. The continuous traffic data (e.g., traffic flow, and speed) from various channels and nodes in a traffic network are coupled with each other over the time points of each channel, spatially between traffic nodes, and jointly in both spatial and temporal dimensions. Such multi-aspect traffic data couplings reflect the conditions of a real-life traffic system and evolve over traffic movement and network dynamics. The recent studies formulate traffic prediction by high-profile graph neural networks. However, they mainly focus on hidden relations captured by neural graph mechanisms while overlooking or simplifying the above multi-aspect traffic data couplings. By modeling a traffic system as a coupled traffic network, we learn the multi-aspect traffic data couplings by a Multi-relational Synchronous Graph Attention Network (MS-GAT). Specifically, MS-GAT learns three embeddings to respectively but synchronously represent the traffic data-based channel, temporal, and spatial relations between nodes by specific graph attention designs. The embeddings are further adaptively coupled according to their respective importance to prediction. Tested on five real-world datasets, MS-GAT outperforms six SOTA graph networks-based traffic predictors. MS-GAT captures not only spatial and temporal couplings but also traffic data-based channel interactions over traffic evolution.

As network and autonomous driving technologies rapidly advance, traffic flow prediction has become a crucial area of research. It plays a significant role in optimizing urban traffic management and enhancing road safety, drawing increasing attention from researchers. As a specific form of time-series data, traffic flow data is often used in prediction tasks utilizing large language models. Recent developments in graph data and improvements in graph neural networks have led researchers to employ methods like adjacency and Laplacian matrices for addressing relational issues among distant nodes. However, most existing methods focus on enhancing prediction performance through network architectures or using adaptive matrices to capture spatiotemporal relationships, with limited exploration into the impact of input embedding. This paper introduces an innovative approach to traffic flow prediction, the ASTRformer, which emphasizes the fusion of spatial and temporal information in historical data through an adaptive spatio-temporal relation learning mechanism. This mechanism integrates feature embedding with adaptive spatial and temporal embeddings. A learnable spatio-temporal fusion network then parameterizes these embeddings, producing the input representations. Subsequently, a transformer model captures these representations to predict future traffic flows. Experimental results on six datasets demonstrate that our method effectively captures spatio-temporal dependencies, achieving state-of-the-art performance across various prediction metrics.

Trajectory prediction in autonomous driving relies on accurate representation of all relevant contexts of the driving scene, including traffic participants, road topology, traffic signs, as well as their semantic relations to each other. Despite increased attention to this issue, most approaches in trajectory prediction do not consider all of these factors sufficiently. We present SemanticFormer, an approach for predicting multimodal trajectories by reasoning over a semantic traffic scene graph using a hybrid approach. It utilizes high-level information in the form of meta-paths, i.e. trajectories on which an agent is allowed to drive from a knowledge graph which is then processed by a novel pipeline based on multiple attention mechanisms to predict accurate trajectories. SemanticFormer comprises a hierarchical heterogeneous graph encoder to capture spatio-temporal and relational information across agents as well as between agents and road elements. Further, it includes a predictor to fuse different encodings and decode trajectories with probabilities. Finally, a refinement module assesses permitted meta-paths of trajectories and speed profiles to obtain final predicted trajectories. Evaluation of the nuScenes benchmark demonstrates improved performance compared to several SOTA methods. In addition, we demonstrate that our knowledge graph can be easily added to two graph-based existing SOTA methods, namely VectorNet and LaFormer, replacing their original homogeneous graphs. The evaluation results suggest that by adding our knowledge graph the performance of the original methods is enhanced by 5% and 4%, respectively.

Having knowledge of neighboring agents' motion patterns along with collision probability is a key challenge for heterogeneous agents' trajectory prediction. In this article, we present rich relational feature learning for efficacious and safe traffic agents' motion forecast. The global temporal information is leveraged using co-attention in feature space. The proposed model not only considers observed trajectories and agents' relational patterns but also the model is learned to be conscious of collision likelihood. The extent of collision likelihood is computed for each agent's move and guided to long short-term memory network during model training. Ground-truth information about the collision alertness among neighboring nodes' trajectories is not available, therefore, reinforcement learning is employed for learning this task. Extensive evaluation results on Apolloscape and Argoverse benchmark datasets are conducted. A substantial performance improvement of the proposed method over the state-of-the-art methods is achieved in terms of average displacement error and the final displacement error.

For automotive applications, the Graph Attention Network (GAT) is a prominently used architecture to include relational information of a traffic scenario during feature embedding. As shown in this work, however, one of the most popular GAT realizations, namely GATv2, has potential pitfalls that hinder an optimal parameter learning. Especially for small and sparse graph structures a proper optimization is problematic. To surpass limitations, this work proposes architectural modifications of GATv2. In controlled experiments, it is shown that the proposed model adaptions improve prediction performance in a node-level regression task and make it more robust to parameter initialization. This work aims for a better understanding of the attention mechanism and analyzes its interpretability of identifying causal importance.

Online ride-hailing services have become an important component of urban transportation in recent years. As a fundamental research problem for such services, the timely prediction of passenger demands in different regions is vital for effective traffic flow control. As both spatial and temporal patterns are indispensable passenger demand prediction, relevant research has evolved from pure time series to graph-structured data for modelling historical passenger demand data, where a snapshot graph is constructed for each time slot by connecting region nodes via different relational edges. Consequently, the spatiotemporal passenger demand records naturally carry dynamic patterns in the constructed graphs, where the edges also encode important information about the directions and volume (i.e., weights) of passenger demands between two connected regions. However, existing graph-based solutions fail to simultaneously consider those three crucial aspects of dynamic, directed and weighted (DDW) graphs, leading to limited expressiveness when learning graph representations for passenger demand prediction. Therefore, we propose a novel spatiotemporal graph attention network, namely Gallat (Graph prediction with all attention) as a solution. In Gallat, by comprehensively incorporating those three intrinsic properties of DDW graphs, we build three attention layers to fully capture the spatiotemporal dependencies among different regions across all historical time slots. Our experimental results on real-world datasets demonstrate that Gallat outperforms the state-of-the-art approaches.

Travel time estimation (TTE) on a specific route is a challenging task since the complex road network structure and hard-captured temporal patterns. Many excellent methods have been proposed to address the aforementioned problems. Some approaches well designed heuristically in a non-learning based way have the advantage of a quick response to the query for travel time estimation. However, these methods are largely affected by the noise of traffic data since they are limited to a single feature. Existing road segment based methods are generally considered intuitive but are not accurate enough for they fail to model complex factors, like delay and direction of intersections. In this paper, we propose a novel attention based sequence learning model for travel time estimation of a path (ASTTE), that not only considers the real-world road network topology as multi-relational data but also refine the problem to the road segment and intersection direction aspects. Besides, we integrate the traffic information as local and neighbor dependency, which helps to monitor dynamic traffic conditions during the trip. The use of the attention mechanism allows the model to focus on significant elements among the path comprises road segments and intersections. High-quality experiments on two real-world datasets have demonstrated the effectiveness and robustness of our framework.

With the wide application of vehicular location-based services, precise estimation of the travel time plays a crucial role in intelligent transportation systems, such as driving navigation, traffic monitoring and route planning. Recent methods have made significant progress on public datasets, but are not satisfied for current ride-hailing platforms with complex road network topology and dynamic traffic fluctuation. In this paper, we propose an end-to-end Deep Fusion framework for Travel Time Estimation (DFTTE), which exploits multi-source heterogeneous traffic information within an Encoder-Decoder architecture. Specifically, we explore a relational fusion network to learn the relationship of road link segments, and employ an attention mechanism to capture efficient correlations among spatial and temporal features. Extensive experiments have been conducted on two large-scale real-world traffic datasets collected by DiDi platform, and the results have demonstrated our effectiveness compared with the state-of-the-art.

In the context of rapidly growing city road networks, understanding complex traffic patterns and implementing effective safety monitoring through advanced Transportation Cyber-Physical Systems (T-CPS) has become increasingly challenging. This involves understanding spatial relationships and non-linear temporal associations. Accurately predicting traffic in such scenarios, particularly for long-term sequences, is challenging due to the complexity of the data. Traditional ways of predicting traffic flow use a single fixed graph structure based on location. This structure does not consider possible correlations and cannot fully capture long-term temporal relationships among traffic flow data, thereby limiting the system ability to ensure safety and reliability. To address this challenge, we propose a novel traffic prediction framework called Attention-based Spatio-temporal Multi-scale Graph Convolutional Recurrent Network (ASTMGCNet). This study introduces a novel framework designed to improve prediction accuracy in dynamic urban traffic systems by effectively capturing complex spatio-temporal correlations through multi-scale feature extraction and attention mechanisms. ASTMGCNet records changing features of space and time by combining Gated Recurrent Units (GRU) and Graph Convolutional Networks (GCN). Its design incorporates multi-scale feature extraction and dual attention mechanisms, effectively capturing informative patterns at different levels of detail. This strategic design allows ASTMGCNet to effectively capture complex spatio-temporal correlations within traffic sequences, enhancing prediction accuracy. We have tested this method on two different real-world datasets and found that ASTMGCNet predicts significantly better than other methods, demonstrating its potential to advance traffic flow prediction and improve safety and reliability in T-CPS applications.

Solving traffic assignment problem for large networks is computationally challenging when conventional optimization-based methods are used. In our research, we develop an innovative surrogate model for a traffic assignment when multi-class vehicles are involved. We do so by employing heterogeneous graph neural networks which use a multiple-view graph attention mechanism tailored to different vehicle classes, along with additional links connecting origin-destination pairs. We also integrate the node-based flow conservation law into the loss function. As a result, our model adheres to flow conservation while delivering highly accurate predictions for link flows and utilization ratios. Through numerical experiments conducted on urban transportation networks, we demonstrate that our model surpasses traditional neural network approaches in convergence speed and predictive accuracy in both user equilibrium and system optimal versions of traffic assignment.

Traffic Prediction based on graph structures is a challenging task given that road networks are typically complex structures and the data to be analyzed contains variable temporal features. Further, the quality of the spatial feature extraction is highly dependent on the weight settings of the graph structures. In the transportation field, the weights of these graph structures are currently calculated based on factors like the distance between roads. However, these methods do not take into account the characteristics of the road itself or the correlations between different traffic flows. Existing approaches usually pay more attention to local spatial dependencies extraction while global spatial dependencies are ignored. Another major problem is how to extract sufficient information at limited depth of graph structures. To address these challenges, we propose a Random Graph Diffusion Attention Network (RGDAN) for traffic prediction. RGDAN comprises a graph diffusion attention module and a temporal attention module. The graph diffusion attention module can adjust its weights by learning from data like a CNN to capture more realistic spatial dependencies. The temporal attention module captures the temporal correlations. Experiments on three large-scale public datasets demonstrate that RGDAN produces predictions with 2%-5% more precision than state-of-the-art methods.

In recent years, with the in-depth application of AI technology in the field of spatiotemporal fusion, modeling of complex spatiotemporal dependencies of small sample data has become a hot topic. However, existing methods usually have three major shortcomings: first, the spatial dependency based on the static adjacency matrix is difficult to reflect the real-time evolution of the road network; second, the spatiotemporal features are often modeled separately, and deep coupling cannot be achieved; third, the model structure complexity is high, which does not match the scale of small sample data and easily leads to overfitting. To this end, this paper proposes LCA-STGNet (Link-Coefficient Adaptation Spatio-Temporal GNN), which uses LSTM as a temporal encoder, uses a dynamic adjacency matrix to capture time-varying spatial features, and enhances information interaction and expression capabilities through cross-node attention mechanisms and residual connections. On a multi-section traffic speed dataset, LCA-STGNet performs significantly better than HA, ARIMA, LSTM, GCN+LSTM, and STGCN in multi-step prediction: compared with LSTM, the average RMSE is reduced by $41.5 \%$; compared with STGCN, MAE is reduced by $10.1 \%$ and $\mathbf{R}^{\mathbf{2}}$ is increased by 4.3 percentage points; and it remains robust in long sequence predictions, verifying its robustness and generalization ability in small sample scenarios.

Modeling complex spatial and temporal dependencies in multivariate time series data is crucial for traffic forecasting. Graph convolutional networks have proved to be effective in predicting multivariate time series. Although a predefined graph structure can help the model converge to good results quickly, it also limits the further improvement of the model due to its stationary state. In addition, current methods may not converge on some datasets due to the graph structure of these datasets being difficult to learn. Motivated by this, we propose a novel model named Dynamic Correlation Graph Convolutional Network (DCGCN) in this paper. The model can construct adjacency matrices from input data using a correlation coefficient; thus, dynamic correlation graph convolution is used for capturing spatial dependencies. Meanwhile, gated temporal convolution is used for modeling temporal dependencies. Finally, we performed extensive experiments to evaluate the performance of our proposed method against ten existing well-recognized baseline methods using two original and four public datasets.

Traffic flow prediction is fundamental to the dynamic control and application of Intelligent Transportation Systems (ITS), which play a crucial role in alleviating road congestion. However, existing approaches have not fully exploited the inherent dynamic and multifaceted spatiotemporal features within traffic data, posing significant challenges in achieving accurate traffic flow predictions. To address this issue, we propose a novel Multi Dynamic Temporal Representation Graph Convolutional Network (MDTRGCN). Specifically, we introduce a dynamic graph construction method that learns the time‒space dependencies specific to road segments. On the basis of this method, we develop a dynamic graph convolution module that aggregates the hidden states of neighboring nodes to a focal node by propagating messages across a dynamic adjacency matrix. Moreover, a multiaspect fusion module is presented, which combines auxiliary hidden states learned from traffic volume with primary hidden states derived from traffic speed. Finally, we propose a temporal representation module that infers the content of masked subsequences from small portions of unmasked subsequences and their temporal context. The experimental results on real-world datasets demonstrate that the proposed method not only achieves state-of-the-art predictive performance but also provides clear and interpretable insights into the dynamic spatial relationships of road segments.

Traffic prediction plays an essential role in intelligent transportation systems by supporting urban traffic management and public safety. A major challenge lies in addressing both the limitations of static assumptions and the inherent complexity they introduce when modeling dynamic and heterogeneous traffic systems. Traditional methods often simplify complex spatio-temporal data into a single-dimensional framework, potentially overlooking intricate node interactions and detailed network characteristics. This fundamental challenge manifests primarily in single-task approaches. When extended to multi-task learning scenarios, the complexity and limitations of this modeling challenge becomes more pronounced. To address these issues, this paper introduce a novel framework, Dynamic Graph Transformation with Multi-Task Learning (DGT-MTL) for spatio-temporal traffic prediction. DGT-MTL features a dynamic adjacency matrix generation module that balances static stability with dynamic flexibility. Additionally, it employs a multi-scale graph learning module to effectively capture fine-grained, latent features. An adaptive multi-task learning module is incorporated to uncover hidden correlations and dynamic relationships between road segments. Experiments conducted across six standard benchmarks demonstrate DGT-MTL's superior performance compared to contemporary approaches, achieving over 15 % improvements in both ROC-AUC and F1 score metrics. Further experiments demonstrate its effectiveness and robustness in handling complex traffic prediction.

Neural network models based on GNNs often achieve good results in traffic flow prediction tasks of traffic networks. However, most existing GNN-based methods apply a fixed graph structure to capture spatial dependencies between nodes, and fixed graph structures may not be able to reflect the spatiotemporal changes in node dependencies. To address this, introducing a self-attention mechanism applied to an adaptive adjacency matrix, the neural network architecture is improved based on Graph WaveNet, and a new approach called self-attention dynamic graph wave network (SA-DGWN) is proposed, which can fit the spatiotemporal dependencies of the road network. In an experiment, traffic flow data were extracted based on RFID from certain roads in Nanjing, China. The results show that under the same configuration, compared to Graph WaveNet, MAE, MAPE, and RMSE from the proposed method reduced by 3.08%, 3.68%, and 2.6%, respectively. In addition, for the training data, we explored the impact of temporal feature and sampling periods on the training effect. The additional results indicate that adding hour-minute-second information to the input improved the model’s accuracy, reducing MAE, MAPE, and RMSE by 15.28%, 12.28%, and 14.01%, respectively. Adding day-of-the-week features also brought substantial performance improvements. For different sampling periods, the model performed better overall with a 10 min sampling period compared to 5 min and 15 min periods. For single-step prediction tasks, the longer the sampling period, the better the prediction effect.

Accurate traffic flow prediction is crucial for alleviating traffic congestion and optimizing intelligent transportation system (ITS). However, traffic flow is subject to uncertainties and exhibits complex spatial and temporal dependence and dynamic change characteristics. Moreover, many efforts rely on a single view, which makes it difficult to comprehensively capture multiple levels of spatial and temporal correlations, thus limiting the accuracy of predictions. Therefore, we propose the multiview spatiotemporal dynamic graph convolution framework MVSTDG for more comprehensively exploring and fusing the multiview spatiotemporal features. First, we design a dual-path time-patch convolution module (TPConv) module to separately model short-term fluctuations and long-term periodic trends, enabling effective extraction of dynamic features at multiple temporal scales. Second, we construct a data-driven traffic pattern library to generate dynamic adjacency matrices and integrate them with static topologies view. An adaptive diffusion graph convolutional network (ADGCN) is then employed to model both local and global spatial correlations. In addition, we design a cross-gated spatiotemporal fusion mechanism that adaptively adjusts the contribution of short-term and long-term information, enhances the interaction of spatiotemporal information, and improves the model’s adaptive capability under different time scales. The experimental results show that MVSTDG outperforms the state-of-the-art baselines in several evaluation metrics and demonstrates higher prediction accuracy and stability on the four real datasets.

As urbanization accelerates and vehicle volumes grow, turn-level queue length prediction at signalized intersections becomes crucial for traffic management. This study proposes a dynamic graph deep learning model that incorporates traffic signal states to predict maximum turn-level queue lengths. A simulation-based dataset was created by varying network structure, traffic demand, and signal plans. The data were represented as dynamic graphs with time-varying adjacency matrices based on traffic signal states. A Graph Attention Network (GAT) was used to capture spatial dependencies, followed by a Gated Recurrent Unit (GRU) for temporal modeling. A congestion-aware weighted loss function was introduced to improve accuracy under heavy traffic. The model's performance was evaluated against various baselines and signal representations using MAE and RMSE. Results show that the proposed model outperforms existing methods, even in extended prediction horizons. These findings highlight the potential of dynamic graph learning for detailed turn-level traffic prediction for traffic management.

Traffic flow prediction presents significant challenges due to complex spatio-temporal dependencies. Conventional static road network models fail to adequately capture dynamic traffic patterns and struggle with multi-scale feature extraction, limiting prediction accuracy. To address these problems, we present DMAGCN (Dynamic Multi-scale Adaptive Graph Convolutional Network), an innovative architecture combining MGTCN (Multi-scale Gated Temporal Convolution Network) and ADMGCN (Adaptive Dynamic Multi-Graph Convolutional Network) Modules. MGTCN extracts multi-scale temporal features by combining temporal attention mechanisms with gated convolutional networks, while ADMGCN enhances spatial representation through graph convolutions with spatial attention layers and adaptive adjacency matrices. Comprehensive experimental evaluations conducted on the PEMS04 and PEMS08 datasets demonstrate that DMAGCN consistently outperforms existing state-of-the-art methods in traffic flow prediction tasks.

Accurate traffic flow prediction is essential for the effective management of Intelligent Transportation Systems (ITS). However, traditional methods based on static graph structures often fail to address the complex and nonlinear spatiotemporal dependencies in evolving traffic conditions. To address this challenge, we propose a Dynamic Graph Convolutional Recurrent Network with Temporal Self‐Attention (DGCRN‐TSA), which integrates a temporal attention mechanism to jointly capture dynamic spatial topologies and long‐range temporal patterns. The model incorporates a graph generation module that adaptively learns time‐varying adjacency matrices from traffic signals and introduces a trend‐aware attention module enhanced by residual‐guided decomposition for distinguishing between normal and anomalous traffic behaviours. Experiments on real traffic datasets confirm that DGCRN‐TSA achieves superior performance in both short‐ and medium‐to‐long‐term forecasts. Notably, it reduces MAE by 19.4% on PeMS04 and improves MAPE by 12.2% on PeMS08. The model also ensures high prediction accuracy with strong computational efficiency and an inference speed comparable to AGCRN. DGCRN‐TSA offers an efficient and reliable solution for dynamic spatiotemporal modelling and large‐scale real‐time traffic prediction.

No abstract available

As intelligent transportation systems continue to develop, the real-time and accurate traffic prediction in urban road networks becomes increasingly critical. Urban road network topology frequently changes due to traffic light control. To capture the dynamic spatial correlations among urban roads, this article introduces a traffic forecasting method utilizing a dynamic spatial-temporal graph convolutional network (D-TGCN) model integrated with an attention mechanism. The method incorporates the traffic BERT model to enhance the static adjacency matrix in traditional graph convolutional neural networks (CNNs) by leveraging the attention mechanism, thereby capturing the implicit correlation between the dynamic variations in the urban road network and traffic flow. First, the road network is transformed into dynamic graph sequences, which traffic BERT uses to generate the final dynamic correlation matrix. Subsequently, the graph convolutional network (GCN) is employed alongside the dynamic correlation matrix to capture dynamic spatial dependencies, while temporal dependencies are modeled using a gated recurrent unit (GRU). LOS-loop and SZ-taxi are the two real-world traffic datasets that were used to test and validate the enhanced model. Results from the experiments show that the D-TGCN model outperformed the temporal graph convolutional network (T-GCN) model by 11.08%, 12.23%, 13.05%, and 13.71% in prediction tasks spanning 15, 20, 30, 45, and 60 min. These results show that the D-TGCN model gives considerable benefits for long-term forecasting and delivers improved prediction accuracy.

With increasingly complicated urban transportation networks, precise traffic flow prediction provides a solid foundation for intelligent traffic management. This paper introduces a Hybrid Graph Memory Network (HGMN), a novel approach designed to model and forecast complex spatio-temporal dependencies in traffic dynamics. We employ one-dimensional convolutional layers to extract hierarchical features from inputs, facilitating structured representations across various extraction levels through pooling layers. Then, we introduce a graph memory network, which incorporates the decomposed features into a dynamic memory graph. This approach integrates adjacency matrices with time-varying features to capture the spatio-temporal dependencies inherent within the data. Finally, an upsampling mechanism fuses features from different hierarchical levels, ensuring the effective integration of detailed information and promoting the effective memory retention and prediction of short- and long-term traffic dynamics. Experiments across four distinct datasets are conducted to validate the performance of HGMN, alongside a sensitivity analysis to examine the impact of different components. The findings indicate that, compared to the benchmark, HGMN achieves a significant reduction in model sizes by 46.1%, leading to decreased memory usage, while still maintaining an accuracy improvement by 5.7%.

Accurate and reliable short-term traffic forecasting is critical for modern intelligent transportation systems. This paper introduces a novel hybrid architecture, GDSTGCN, which leverages a dynamic graph constructor and spatio-temporal graph convolutions, and presents a rigorous comparative study of two temporal processing modules: Temporal Convolutional Networks (TCN) and Long Short-Term Memory (LSTM) networks. The framework dynamically builds generalized adjacency matrices from input sequences and utilizes Chebyshev spectral graph convolutions to capture complex spatio-temporal dependencies. Both model variants were evaluated on the widely-used METR-LA benchmark dataset for a 60-minute forecast horizon. The results demonstrate the superiority of the convolutional approach, with the GDSTGCN-TCN model achieving a highly competitive Mean Absolute Error (MAE) of 3.48 and a Mean Absolute Percentage Error (MAPE) of 9.8%. This performance surpassed the LSTM variant and validates the efficacy of combining dynamic graph structures with specialized temporal modules for robust traffic forecasting.

In recent years, systems of intelligent transportation have increasingly relied on advanced machine learning and deep learning models for accurate traffic forecasting, which is critical for urban mobility, and smart city development. Traditional work on Spatio-Temporal Gated Graph Attention Networks (STGGAN) has shown promising results. However, model faces challenges due to its dependence on a static graph structure, which cannot adapt to dynamic traffic variations, and its single-task focus on speed prediction. Therefore, a Dynamic Multi-Task Spatio-Temporal Gated Graph Attention Network (DM-STGGAN) is proposed. Initially, traffic datasets like California Performance Measurement System District 4 (PeMSD4) and PeMSD8 are preprocessed by cleaning missing values and deriving indicators of key traffic including flow, density, and average speed. Further, a Dynamic Graph Learner adaptively updates adjacency matrices using both road topology and real-time traffic correlations followed by node and edge feature representation. Then, temporal dependencies are captured using Gated Recurrent Units (GRUs), while edge-aware multihead attention with a gated mechanism models complex spatial dependencies. Furthermore, a multi-task output layer jointly predicts speed, flow, and density for comprehensive forecasting. Experimental results demonstrate that proposed DM-STGGAN achieves lower root mean square error (2.7 speed), mean absolute error (1.6 speed) and higher accuracy (98.4 %) respectively.

Accurate traffic flow prediction is vital for intelligent transportation systems, yet strong spatiotemporal coupling and multi-scale dynamics make modelling difficult. Existing methods often rely on static adjacency and short input windows, limiting adaptation to time-varying spatial relations and long-term patterns. To address these issues, we propose the Pre-trained Trend-aware Dynamic Graph Convolutional Network (PT-TDGCN), a two-stage framework. In the pre-training stage, a Transformer-based masked autoencoder learns segment-level temporal representations from historical sequences. In the prediction stage, three designs are integrated: (1) dynamic graph learning parameterized by tensor decomposition; (2) convolutional trend-aware attention that adds 1D convolutions to capture local trends while preserving global context; and (3) spatial graph convolution combined with lightweight fusion projection for aligning pre-trained, spatial, and temporal representations. Extensive experiments on four real-world datasets demonstrated that PT-TDGCN consistently outperformed 14 baseline models, achieving superior predictive accuracy and robustness.

The problem of traffic forecasting has received much attention as a central part of intelligent transportation systems. In recent years, many different models have been proposed to improve the performance of traffic forecasting. However, there are some problems with these models: they only focus on the dependencies between nodes and ignore the dependencies between edges; the highly dynamic spatial dependencies of traffic networks in time are not fully considered. In this paper, we propose a multi-scale dynamic spatial-temporal graph convolution network with edge feature embedding(MDSTGCN). In the spatial dimension, we construct the dynamic adjacency matrix and the hypergraph. Capturing spatial correlations using diffusion convolution. In the temporal dimension, we design a multi-scale temporal convolution module to capture the temporal dynamics of traffic data at different scales. We conducted experiments on four real datasets and the results show that our model outperforms the baseline models.

The accurate prediction of traffic conditions is essential for effective and efficient traffic management and control. The dynamic and complex nature of traffic data, characterized by intricate temporal and spatial features, presents significant challenges to accurate traffic forecasting. While previous studies have developed various models with advanced algorithms, they often fail to fully capture the holistic spatio-temporal features and the dynamically evolving correlations within traffic networks. Additionally, these studies often overlook the potential of adjacency matrices learned from real-time traffic data to more accurately represent the interconnectivity of nodes within road network. To address these gaps, this study introduces the Generalized Dynamic Spatio-Temporal Graph Convolutional Network (GDSTGCN), a novel prediction model tailored for traffic data. First, this model builds a learning-based generalized dynamic graph structure, which incorporates both spatial and temporal connections and evolves with real-time traffic data. Then, a generalized dynamic graph convolution, integrated with graph diffusion, is crafted to operate on the designed generalized dynamic graph structure. This plays a critical role in holistically capturing local and global spatio-temporal traffic dependencies. Moreover, the generalized dynamic graph convolution is incorporated with Temporal convolution and other essential components, forming a cohesive framework that enables effective and efficient traffic flow predictions. To validate the performance of the GDSTGCN model, we conducted extensive experiments using four real-world road network datasets. The results demonstrate that our model outperforms existing state-of-the-art GCN-based models and traditional baseline methods.

No abstract available

In the field of intelligent transportation systems, accurately predicting traffic flow is a challenging endeavor for enabling efficient services in urban road networks. While current research has made significant advancements in capturing nonlinear features and spatiotemporal dependencies at a fixed geographic node level, it often overlooks the dynamic interactions of traffic flow across different regions. In this work, we present a novel local-global dynamic information fusion graph learning model (LGDF-GL) for traffic flow prediction. Different from other graph learning model, we first design a dynamic adjacency matrix by region-adaptive parameter learning, and fuse it with a static adjacency matrix based on node geographic information by dynamic information fusion module, and the fused result serves as the diagonal elements of the local adjacency matrix. Subsequently, we input it into the graph convolutional network (GCN) and gated recurrent unit (GRU) aggregation model to extract local feature information containing dynamic spatiotemporal features from traffic flow data. The local feature information is pruned and concatenated to obtain global feature representations. Then we achieve global feature representation prediction through multiple fully connected layers. Finally, we conducted experiments on two real-world traffic flow datasets to evaluate the performance of our proposed method. The results clearly demonstrate that the LGDF-GL method achieves superior prediction ability compared to several baseline models.

It is crucial for both traffic management organisations and individual commuters to be able to forecast traffic flows accurately. Graph neural networks made great strides in this field owing to their exceptional capacity to capture spatial correlations. However, existing approaches predominantly focus on local geographic correlations, ignoring cross-region interdependencies in a global context, which is insufficient to extract comprehensive semantic relationships, thereby limiting prediction accuracy. Additionally, most GCN-based models rely on pre-defined graphs and unchanging adjacency matrices to reflect the spatial relationships among node features, neglecting the dynamics of spatio-temporal features and leading to challenges in capturing the complexity and dynamic spatial dependencies in traffic data. To tackle these issues, this paper puts forward a fresh approach: a new self-supervised dynamic spatio-temporal graph convolutional network (SDSC) for traffic flow forecasting. The proposed SDSC model is a hierarchically structured graph–neural architecture that is intended to augment the representation of dynamic traffic patterns through a self-supervised learning paradigm. Specifically, a dynamic graph is created using a combination of temporal, spatial, and traffic data; then, a regional graph is constructed based on geographic correlation using clustering to capture cross-regional interdependencies. In the feature learning module, spatio-temporal correlations in traffic data are subjected to recursive extraction using dynamic graph convolution facilitated by Recurrent Neural Networks (RNNs). Furthermore, self-supervised learning is embedded within the network training process as an auxiliary task, with the objective of enhancing the prediction task by optimising the mutual information of the learned features across the two graph networks. The superior performance of the proposed SDSC model in comparison with SOTA approaches was confirmed by comprehensive experiments conducted on real road datasets, PeMSD4 and PeMSD8. These findings validate the efficacy of dynamic graph modelling and self-supervision tasks in improving the precision of traffic flow prediction.

Dynamic Graph Neural Networks (DGNNs) have become one of the most promising methods for traffic speed forecasting. However, when adapting DGNNs for traffic speed forecasting, existing approaches are usually built on a static adjacency matrix (no matter predefined or self-learned) to learn spatial relationships among different road segments, even if the impact of two road segments can be changeable dynamically during a day. Moreover, the future traffic speed cannot only be related with the current traffic speed, but also be affected by other factors such as traffic volumes. To this end, in this paper, we aim to explore these dynamic and multi-faceted spatio-temporal characteristics inherent in traffic data for further unleashing the power of DGNNs for better traffic speed forecasting. Specifically, we design a dynamic graph construction method to learn the time-specific spatial dependencies of road segments. Then, a dynamic graph convolution module is proposed to aggregate hidden states of neighbor nodes to focal nodes by message passing on the dynamic adjacency matrices. Moreover, a multi-faceted fusion module is provided to incorporate the auxiliary hidden states learned from traffic volumes with the primary hidden states learned from traffic speeds. Finally, experimental results on real-world data demonstrate that our method can not only achieve the state-of-the-art prediction performances, but also obtain the explicit and interpretable dynamic spatial relationships of road segments.

Accurate traffic flow prediction is crucial for alleviating traffic congestion, reducing environmental pollution, and enhancing travel efficiency. However, due to the complex periodicity and real-time dynamics of traffic data, achieving accurate traffic flow prediction is a significant challenge. Most existing methods use predefined or dynamic adjacency matrices to model spatial dependencies, overlooking the periodic characteristics of traffic data. This paper proposes a novel Dynamic Periodic Graph Attention Recurrent Network (DPGARN) for traffic flow prediction, which simultaneously considers the dynamism and periodicity of traffic data. In DPGARN, we use a dynamic graph generation module to generate a dynamic adjacency matrix for each time step, capturing real-time spatial dependencies. This module uses spatial embedding, daily embedding, and weekly embedding to model dynamism and periodicity. The generated dynamic adjacency matrices are integrated into a Graph Attention Network (GAT), forming the Dynamic Graph Attention Network (D-GAT). We combine bidirectional GRU (Bi-GRU) with D-GAT to learn local Spatio-temporal dependencies at different time steps and use an attention mechanism to capture global Spatio-temporal dependencies. Experimental results on four datasets demonstrate that DPGARN exhibits the best predictive performance.

Traffic flow prediction remains a challenging task due to the complexity of spatio-temporal dependencies and the dynamic nature of traffic patterns. Traditional methods often struggle to efficiently model long-term dependencies and adapt to dynamic changes in traffic networks, leading to suboptimal prediction performance. To address these challenges, this paper proposes a novel traffic flow prediction model that integrates multi-scale convolutional networks and dynamic graph convolutional networks. By replacing the traditional Transformer architecture with Linformer, the model significantly reduces computational complexity while efficiently processing long sequential data. Additionally, the introduction of multi-scale convolutional networks allows for effective extraction of flow features at short, medium, and long time scales, enhancing prediction accuracy across different horizons. The incorporation of Dynamic Graph Convolutional Networks (Dynamic GCN) enables dynamic adjustment of the adjacency matrix, allowing the model to adaptively capture changes in spatio-temporal relationships within the traffic network. Experimental results show that the proposed model outperforms existing approaches in short-, medium-, and long-term traffic flow prediction tasks across multiple datasets, demonstrating its effectiveness in addressing the complexities of spatio-temporal dependencies and dynamic changes.

Spatial-temporal data mining for traffic flow regulation remains a challenging task due to the complex spatiotemporal correlation and the non-Euclidean nature of traffic data. Existing approaches typically utilize a provided spatial adjacency graph and sophisticated mechanisms for modeling spatial and temporal correlations. However, the static representation of a given spatial adjacency graph may restrict effective spatial-temporal dependency learning and fail to capture dynamic interactions over time. To address those challenges, our paper proposes the Dynamic Adaptive Graph Convolution Network (DAGCN), a highly optimized model with a low memory footprint. DAGCN can effectively learn hidden spatial dependencies by utilizing a novel graph convolution layer with adaptive parameter update mechanisms and innovative attention computations. Additionally, DAGCN could overcome the limitations of static data dependencies by employing the dynamic graph modulation approach. This approach combines the assumption of static adjacency matrices with dynamic feature generation based on the multi-head attention mechanism. Meanwhile, by integrating this module with a novel temporal module, DAGCN could gather information in the temporal domain. Experimental results on two real traffic datasets demonstrate that the proposed model achieves state-of-the-art performance compared to other baselines while maintaining the lowest parameter counts.

—Traffic speed prediction based on spatial-temporal data plays an important role in intelligent transportation. The time-varying dynamic spatial relationship and complex spatial-temporal dependence are still important problems to be considered in traffic prediction. In response to existing problems, a Dynamic and Static Graph Convolutional Recurrent Network (DASGCRN) model for traffic speed prediction is proposed to capture the spatial-temporal correlation in the road network. DASGCRN consists of Spatial Correlation Extraction Module (SCEM), Dynamic Graph Construction Module (DGCM), Dynamic Graph Convolution Recurrent Module (DGCRM) and residual decomposition. Firstly, the improved traditional static adjacency matrix captures the relationship between each time step node. Secondly, the graph convolution captures the overall spatial information between the road networks, and the dynamic graph isomorphic network captures the hidden dynamic dependencies between adjacent time series. Thirdly, spatial-temporal correlation of traffic data is captured based on dynamic graph convolution and gated recurrent unit. Finally, the residual mechanism and the phased learning strategy are introduced to enhance the performance of DASGCRN. We conducted extensive experiments on two real-world traffic speed datasets, and the experimental results show that the performance of DASGCRN is significantly better than all baselines.

In intelligent transportation systems (ITS), traffic prediction has become a core issue in the field of artificial intelligence. Accurate traffic prediction is crucial for reducing congestion, improving travel efficiency, and enhancing traffic safety. However, the complex and dynamically changing spatio-temporal dependencies in traffic networks make the prediction task highly challenging. Most existing methods rely on self-attention mechanisms to capture these spatio-temporal dependencies, which often result in high computational complexity. Furthermore, although traffic systems can be modeled as graph structures, graph convolutional network (GCN), which effectively capture correlations between nodes, typically depend on predefined or adaptive adjacency matrices. These approaches struggle to cope with the highly dynamic nature of real-world traffic systems. To address these issues, we propose a spatio-temporal mamba dynamic graph convolutional recurrent neural network (STMAGRN). This model first leverages the mamba framework to extract spatio-temporal features from traffic sequences. The mamba model, through its unique selective state space mechanism and linear time complexity, achieves efficient sequence processing. Next, we design a spatio-temporal memory module (STM), which identifies and extracts traffic pattern features most similar to the input data, and learns intrinsic representative traffic patterns from the data associated with each node. These patterns are then used to dynamically generate graph structures, providing GCN with real-time updated graph data. Finally, we combine GCN with recurrent neural networks (RNNs) to extract both spatial and temporal features simultaneously. We conducted extensive experiments on six real-world datasets, and the results demonstrate that STMAGRN significantly outperforms state-of-the-art methods in traffic prediction tasks.

Accurate and efficient traffic flow prediction is essential for developing smart cities. Traffic flow data exhibits complex spatio-temporal dependencies, and the weights between nodes may change dynamically due to travel patterns and node attributes. However, existing prediction models primarily rely on static adjacency matrices and complex time series models, which limit model performance. We propose an innovative traffic flow prediction method, the Adaptive Dynamic Spatio-temporal Graph Convolutional Network (ADSTGCN), to address this issue. Specifically, we incorporate a multi-head attention mechanism and an adaptive dynamic adjacency matrix to construct two dynamic spatio-temporal extraction modules, which are integrated with Graph Convolutional Networks (GCN) to overcome the limitation of static adjacency matrices in capturing dynamic spatio-temporal correlations between nodes. At the temporal dimension extraction level, we integrate the Mamba to model long-term time series in traffic flow data, effectively extracting relevant temporal information. Extensive comparative experiments are conducted on four real-world public transportation datasets. The results demonstrate that our model achieves the highest prediction accuracy compared to other baseline models, showcasing its significant potential in traffic flow prediction.