Jul 2026· Computers and artificial intelligence· 0 citations· 7 references
TL;DR
Examination of the spatiotemporal evolution of urban traffic congestion in Beijing from a deep learning perspective based on multi-source data suggests that traffic congestion in Beijing displays a pronounced “dual-peak” pattern associated with daily commuting activities.
Abstract
As urbanization continues to reshape large cities, traffic congestion remains a persistent challenge for urban transportation systems. Using Beijing as a case study, this paper examines the spatiotemporal evolution of urban traffic congestion from a deep learning perspective based on multi-source data. The results suggest that traffic congestion in Beijing displays a pronounced “dual-peak” pattern associated with daily commuting activities. Compared with the pre-pandemic period, weekday travel demand has generally recovered and in some cases exceeded previous levels, whereas holiday travel remains relatively subdued, accompanied by increasingly concentrated travel behavior. Among the models considered, Long Short-Term Memory (LSTM) performs particularly well in capturing nonlinear variations and temporal dependencies in traffic flow, leading to improved prediction accuracy. Between 2020 and 2025, the congestion index experienced a trajectory of decline, recovery, and subsequent stabilization, a pattern that appears to be associated with the gradual implementation of intelligent traffic management measures. These findings contribute to a better understanding of recent changes in urban traffic dynamics and may offer useful insights for future traffic planning and governance.
Urban traffic congestion remains a critical challenge for smart city development, particularly in the context of real-time monitoring and prediction for intelligent transportation systems. To address this issue, this study proposes a Digital Twin-based urban traffic prediction framework using a lightweight Diffusion Convolutional Recurrent Neural Network (DCRNN-Lite). The proposed model integrates spatial dependencies among road segments through diffusion convolution and temporal traffic dynamics through recurrent modeling, enabling effective spatiotemporal traffic forecasting with reduced computational complexity. Experiments were conducted on the real-world METR-LA dataset, consisting of traffic speed data from 207 sensors deployed across the Los Angeles highway network. The experimental results demonstrate that DCRNN-Lite achieves stable prediction performance, as reflected by low Mean Absolute Error (MAE) and Root Mean Square Error (RMSE), consistent convergence behavior, and strong correlation between predicted and actual traffic speeds at both node and city-wide levels. Despite its simplified architecture, the model effectively captures local traffic variations and global mobility trends, making it suitable for real-time deployment. The findings indicate that the proposed approach provides a favorable balance between accuracy and efficiency, highlighting its potential as a core component for digital twin-enabled smart cities and metaverse-based urban traffic management systems.
H. Awad· International Journal Resear...· 0 citations
A logic-oriented analysis of urban traffic spatiotemporal data examines three temporal evolution characteristics: periodicity, trend, and abrupt change and extracts a logical framework for traffic condition prediction.
Shuohan Xu· Applied and Computational En...· 0 citations
Urban traffic congestion continues to increase travel time, fuel consumption, and environmental impact in modern cities, and Tbilisi is no exception: traffic intensity at signalized intersections varies sharply depending on time of day, direction, and the workday–weekend cycle. This study proposes an integrated, AI-assisted framework that connects real-time route-level traffic observation, short-term forecasting, and adaptive signal control within a single decision-support pipeline. Traffic data were collected at five-minute intervals for twelve origin–destination route pairs surrounding one critical intersection in Tbilisi using the Google Routes API, with automated acquisition and storage implemented through AWS Lambda, EventBridge, and Amazon S3, producing a structured dataset of 23,556 observations and 75 variables. A LightGBM regression model forecasts short-term delay at two horizons (t+15 and t+30 minutes), while a two-stage multi-class classifier separates normal from congested traffic states and, within the congested subset, distinguishes slow traffic from jam conditions. Forecasts are translated into operational signal-control actions through a rule-based decision layer and evaluated in a SUMO/TraCI microscopic simulation environment against a fixed-time baseline, a max-pressure-style pressure-adaptive controller, a bridge-model hybrid controller, and an experimental Deep Q-Network (DQN) reinforcement-learning controller. Results show that LightGBM achieved R² = 0.82 and R² = 0.77 at the two forecasting horizons, while the two-stage classifier reached 90.6% overall accuracy. In simulation, the pressure-adaptive controller reduced mean waiting time by approximately 33.6% relative to the fixed-time baseline, while the DQN controller achieved an average reduction of approximately 23.3% across multiple random seeds. The proposed framework demonstrates a reproducible, data-driven pathway from real-time observation to adaptive traffic-signal decision-making, without removing engineering judgment from the control loop.
Nikoloz Patatishvili, Besik Tabatadze· Computational and Applied Sc...· 0 citations
Although the forecasting accuracy of both models decreases with longer prediction horizons, PatchTST maintains relatively lower errors and better stability, and the results indicate that the patch-based representation and attention mechanism provide advantages in capturing both local variations and long-term temporal dependencies in urban traffic sequences.
Wentao Hu, Er Zhou· Applied and Computational En...· 0 citations
Urban traffic congestion in India poses significant economic and environmental challenges, necessitating early detection for effective management. This study presents a novel approach to real-time congestion prediction, addressing the need for a mechanism that can adapt to India's diverse traffic patterns while minimizing computational complexity. Unlike previous studies that focused on long-term forecasts, this research proposes a lightweight, resource-efficient model capable of predicting congestion levels at 15-minute intervals. The model utilizes a real-time dataset collected from the Palakkad region in southern India. To improve prediction accuracy, the method employs ensemble methods. Experiments have shown that the random forest algorithm is 98.6% accurate at predicting traffic jams, which is better than previous research in this area. Additionally, the study develops a best-route recommendation system based on the experimental results. This research contributes to the field by offering a practical approach to mitigating urban traffic congestion in India, with applications in other cities characterized by diverse traffic patterns. The proposed model's ability to provide accurate short-term predictions while maintaining computational efficiency represents a significant advancement in traffic management strategies for developing urban ecosystems. Furthermore, the study underscores the efficacy of data-driven decision support systems in optimizing urban mobility and reducing vehicular emissions.
Sreejith Karunakaran, S. Mathi, Preeja Pradeep· Bulletin of Electrical Engin...· 0 citations
The paper is a detailed research of how big data analytics have been used to predict traffic flow and analyses sources of data, analytics, machine learning and deep- learning models and scalable processing frameworks in modern traffic prediction systems.
Unknown authors· International Journal of App...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
