Architecting Digital Twins for Intelligent Transportation Systems

TL;DR

Digital Twins (DTs) offer real-time monitoring, predictive analytics, and adaptive control for ITS, but existing frameworks struggle with scale, dynamism, and interoperability. This work presents DigIT, a modular DT platform for ITS based on a Domain Concept Model (DCM) that unifies vehicles, sensors, networks, and user behavior, enabling seamless data fusion, prediction, and simulation. The architecture integrates LSTM/BiLSTM traffic forecasting with an adaptive MLOps pipeline, a SUMO-based simulator, and a code-generated interface between models and actuators, all orchestrated by a cloud-edge data lake. Evaluations show accurate short-term predictions ($RMSE$, $MAE$, $MAPE$) and fast runtimes ($ ext{~}7$ ms for predictions; SUMO simulations in $ ext{~}15$ s), validating computational efficiency and fidelity. The framework establishes a scalable path toward real-world ITS deployments, with future work expanding to multi-modal transportation and dynamic calibration.

Abstract

Modern transportation systems face growing challenges in managing traffic flow, ensuring safety, and maintaining operational efficiency amid dynamic traffic patterns. Addressing these challenges requires intelligent solutions capable of real-time monitoring, predictive analytics, and adaptive control. This paper proposes an architecture for DigIT, a Digital Twin (DT) platform for Intelligent Transportation Systems (ITS), designed to overcome the limitations of existing frameworks by offering a modular and scalable solution for traffic management. Built on a Domain Concept Model (DCM), the architecture systematically models key ITS components enabling seamless integration of predictive modeling and simulations. The architecture leverages machine learning models to forecast traffic patterns based on historical and real-time data. To adapt to evolving traffic patterns, the architecture incorporates adaptive Machine Learning Operations (MLOps), automating the deployment and lifecycle management of predictive models. Evaluation results highlight the effectiveness of the architecture in delivering accurate predictions and computational efficiency.

Figures (5)

• Figure 1: Sociotechnical DT Specification Meta Language Concepts for ITS.
• Figure 2: Architecture of $\mathcal{D}ig{IT}$
• Figure 3: Traffic Prediction Dashboard displaying accuracy, prediction errors, and performance metrics.
• Figure 4: Predicted vs Actual Traffic Flow.
• Figure 5: Simulation Dashboard displaying traffic flow timelines, speed distributions, and junction traffic visualizations.