TGDT: A Temporal Graph-based Digital Twin for Urban Traffic Corridors

TL;DR

TGDT addresses urban arterial congestion by integrating Temporal CNNs with Attentional Graph Neural Networks into a modular digital twin for corridors. The model performs interval-based inflow imputation, corridor travel-time estimation, and intersection-level queue and waiting-time predictions through a sequential optimization training regime. A large-scale SUMO-based dataset for Florida’s SR 436 (Real-TMC-like) with static and dynamic graphs enables robust evaluation, where TGDT shows low errors and strong distributional alignment relative to ground truth. Practically, TGDT enables thousands of scenario evaluations in seconds, offering a scalable, interpretable tool for real-time adaptive signal control and urban planning.

Abstract

Urban congestion at signalized intersections leads to significant delays, economic losses, and increased emissions. Existing deep learning models often lack spatial generalizability, rely on complex architectures, and struggle with real-time deployment. To address these limitations, we propose the Temporal Graph-based Digital Twin (TGDT), a scalable framework that integrates Temporal Convolutional Networks and Attentional Graph Neural Networks for dynamic, direction-aware traffic modeling and assessment at urban corridors. TGDT estimates key Measures of Effectiveness (MOEs) for traffic flow optimization at both the intersection level (e.g., queue length, waiting time) and the corridor level (e.g., traffic volume, travel time). Its modular architecture and sequential optimization scheme enable easy extension to any number of intersections and MOEs. The model outperforms state-of-the-art baselines by accurately producing high-dimensional, concurrent multi-output estimates. It also demonstrates high robustness and accuracy across diverse traffic conditions, including extreme scenarios, while relying on only a minimal set of traffic features. Fully parallelized, TGDT can simulate over a thousand scenarios within a matter of seconds, offering a cost-effective, interpretable, and real-time solution for urban traffic management and optimization.

Figures (4)

• Figure 1: The inputs and outputs of the proposed urban corridor digital twin. TGDT takes input parameters (highlighted in orange), including ingress aggregated traffic waveforms, signal timing parameters (e.g., cycle length, offset, and maximum green duration for each phase), driving behavior parameters (e.g., speed, acceleration, space cushion, lane-changing behavior), turning movement ratios, and the distances between intersections along the major corridor. It simultaneously generates multiple outputs (highlighted in blue), such as westbound travel times along the corridor, queue lengths for each lane group phase, and average waiting times for each lane group phase. The time intervals of the output time series match those of the input inflow waveforms.
• Figure 2: Overview of TGDT framework. This diagram illustrates the architecture of our proposed Digital Twin for urban corridors, which consists of three main modules. Simulation records, extracted from the logs of a microscopic traffic simulator, are transformed into graph-structured data that uniquely represent the corridor’s traffic state for each scenario. The inflow module ($M_{\text{inf}}$) performs a graph imputation task to reconstruct 2D traffic volumes on every intermediate road segment. The travel time module ($M_{\text{tt}}$) carries out a graph regression task to estimate bidirectional corridor-level travel time series. Finally, the queue length ($M_{\text{ql}}$) and waiting time ($M_{\text{wt}}$) modules apply temporal convolution and deconvolution operations on the spatiotemporal representations learned by $M_{\text{tt}}$, producing 3D outputs for maximum queue length and waiting time. These estimates are generated at the intersection and phase levels for each lane group associated with a specific movement phase.
• Figure 3: Visualization outputs of TGDT for a randomly selected traffic scenario. Comparison of actual (red) and predicted (green) curves for several Measures of Effectiveness (MOEs). TGDT takes ingress (inflow) traffic volumes and various parameters, such as traffic signal plans and driving behaviors, to accurately simulate bidirectional travel times throughout urban traffic corridors. It also estimates multi-directional maximum queue lengths and waiting times within a given time interval. The model first imputes intermediate inflow traffic volumes between intersections, then analyzes traffic performance across all intersections ($J_1, ..., J_8$) and all phase lane groups ($P_1, ..., P_8$) simultaneously.
• Figure 4: Predicted versus actual travel time distributions. This figure shows the Kernel Density Estimates (left column) and the joint distribution of travel time characteristics (right column) across the target corridor at a certain time step $T=15$ min. The results suggest that the predicted behavior closely matches the actual travel time in terms of both the average ($\mu$, slightly underestimated) and variability ($\sigma$, slightly overestimated).