Author: Sakshi Haware
Roll no. 65
Introduction
Urban traffic congestion has become one of the biggest challenges facing modern cities. Rapid urbanization, rising private vehicle ownership, and changing work patterns like telecommuting have all contributed to overcrowded roads and longer travel delays. Studies from different regions show that congestion is not only an issue of too many vehicles but also of poor planning, weak infrastructure, and limited use of sustainable transport options. Researchers have examined case studies in cities such as Chennai, Shimla, and Peshawar, as well as advanced models using IoT, VANET, and AI-based forecasting, to understand the causes and possible solutions. Together, these insights highlight that traffic congestion is a complex problem requiring both smart technology and sustainable urban policies.
Literature Review
Telework-induced Urban Sprawl and Traffic Congestion
Gökhan Güven investigates why traffic congestion has returned to pre-pandemic levels despite the permanence of remote work. He argues that telecommuting creates a “rebound effect,” where the time saved from not commuting is replaced by increased travel for leisure, errands, and personal activities. This flexibility also encourages urban sprawl, as employees move away from dense city centers to suburban or rural areas, increasing reliance on private vehicles. The study introduces a mathematical model to derive the social discount rate which governments use to evaluate long-term investments. Güven finds that productivity gains from telework raise SDR, making capital-intensive projects like road expansion appear more attractive. Policymakers therefore prioritize supply-side strategies such as highway construction over demand-side restrictions. By calibrating this model with U.S. data, Güven shows that increased government funding for road projects is a rational response to workforce decentralization. The paper concludes that while telework offers personal flexibility, its macro-level consequences necessitate a shift in how urban planners manage congestion and residential mobility (Güven, 2025).
Assessing Urban Traffic Congestion in Chennai
Sridhar Parkavi and Angamuthu Parthiban focus on two critical intersections in Chennai—Kathipara and T. Nagar. Using video analysis and manual traffic counts, they found both intersections operating at level of service with saturation levels between 1.32 and 1.45. This indicates near-total breakdown of traffic flow. With volumes exceeding 10,000 vehicles per hour, delays surpassed 212 vehicle hours, causing massive economic losses, fuel wastage, and pollution. The study attributes congestion to rapid urbanization and rising private vehicle ownership, which have outpaced infrastructure capacity. The authors recommend adaptive signal control, multimodal corridors, and performance-based management models to alleviate bottlenecks. Their findings emphasize that traditional road expansion is insufficient; instead, sustainable, data-driven strategies are essential to restore mobility and productivity in megacities like Chennai (Parkavi & Parthiban, 2025).
Dynamic Risk Field (DRF) Model
Yuman Xu and Yunxia Li present a Dynamic Risk Field (DRF) model that treats congestion as a physical field, where vehicles and bottlenecks act as source points. Using Gaussian kernel functions, the model predicts how congestion propagates across networks. Unlike traditional “black-box” deep learning models, DRF offers transparency and interpretability, making it more useful for policymakers. By fusing GPS trajectories, road sensors, and weather data, the DRF framework significantly outperforms models like LSTM and GRU, especially in predicting sudden congestion transitions. The study also provides intuitive three-dimensional visualizations of traffic risk, enabling planners to see congestion intensity as heat maps. Applied to Changchun, China, the model demonstrated superior accuracy and scalability. Xu and Li’s work bridges physical theory with AI, offering a robust tool for smart city traffic forecasting and management (Xu & Li, 2025).
Sustainable Mobility Framework
Elena Rossi and Marco Bianchi propose a Congestion Severity Index that integrates idle time, emissions, and public transit reliability to assess congestion hotspots. Their European case studies reveal that fragmented data often leads to reactive policies like road widening, which fail to address sustainability. By adopting CSI, cities can prioritize interventions that balance economic efficiency with environmental responsibility. For example, dedicated bus lanes, cycling infrastructure, and active mobility corridors were shown to reduce congestion more effectively than expanding road capacity. The framework also incorporates predictive machine learning to forecast how infrastructure changes will affect congestion over time. Rossi and Bianchi argue that shifting from a “fix-the-bottleneck” mentality to proactive, sustainability-focused strategies is vital for reducing urban environmental footprints and improving residents’ quality of life (Rossi & Bianchi, 2025).
IoT and Weather-Enhanced Prediction
This study introduces a multi-step ML framework that leverages IoT sensors and weather data to forecast congestion. Ensemble Tree-Based regressors, particularly LightGBM, consistently outperform deep learning models for short-term predictions. By integrating weather as a secondary data stream, the model anticipates congestion influenced by rainfall, temperature, or storms. This approach is especially valuable in cities with limited historical traffic data, where traditional models struggle. The study demonstrates that IoT-enabled forecasting can enhance Intelligent Transportation Systems (ITS), allowing authorities to implement proactive mitigation strategies. Ultimately, this framework provides a practical tool for smart city management, enabling more efficient scheduling and reducing travel delays (Tsalikidis et al., 2024).
VANET Data for Prediction
Wilson Chango and colleagues apply LSTM neural networks to Vehicular Ad-Hoc Network (VANET) data, achieving high accuracy (0.9463). Their model captures sequential traffic dependencies, enabling reliable real-time forecasting. Their dataset included vehicular speed, flow, and weather conditions, allowing for robust predictions. The study highlights that LSTM networks outperform Transformer models in capturing spatiotemporal dependencies. By integrating these predictions into simulation environments, the framework provides scalable solutions for intelligent traffic management systems. The authors conclude that VANET-based insights are essential for optimizing infrastructure and supporting sustainable urban planning (Chango et al., 2025).
Spatiotemporal Hierarchical Analysis
This study introduces a nine-cell grid model to calculate road carrying capacity and identify imbalances between demand and infrastructure. Applied in Chengdu, the framework integrates GIS, remote sensing, and deep learning to reveal spatial heterogeneity in congestion. By analyzing road density, intersection proximity, and vehicle demand, the model identifies unbalanced areas requiring intervention. The hierarchical approach allows precise local adjustments, improving productivity and reducing environmental impacts. The study demonstrates that congestion is not uniform but varies across spatial grids, requiring targeted solutions rather than blanket policies. This method provides a scalable diagnostic tool for traffic managers to balance infrastructure and demand effectively (Jiang et al., 2024).
Mixed Flow of Human-Driven and Autonomous EVs
Chenn-Jung Huang and colleagues address the complexity of mixed traffic environments containing both human-driven and autonomous EVs. Their distributed mechanism uses Support Vector Regression (SVR) to predict traffic flow and optimize routing and charging schedules simultaneously. By coordinating EV fleets and human drivers, the system prevents the “herd effect” where vehicles converge on the same path or charging station. Validated through simulations in Taiwan, the mechanism improved average speeds and reduced energy consumption. The study emphasizes that integrating EVs into urban traffic requires systemic coordination, not isolated solutions (Huang et al., 2021).
Entropy-TOPSIS Method in Peshawar
Daniyal Hussain and colleagues apply a multi-criteria decision-making framework combining Shannon Entropy and TOPSIS to rank congestion hotspots in Peshawar. Using ArcGIS and Passenger Car Unit (PCU) counts, they identified Amin Hotel, PC, and Jalil Kabab as critical bottlenecks, all operating at LOS F. The authors argue that traditional road widening is ineffective and instead recommend integrated management strategies such as “No Parking” zones and smart signaling. Their reproducible model provides urban planners with a data-driven tool to prioritize interventions (Hussain et al., 2026).
Shimla’s Mountain City Challenges
Ajitesh Singh Chandel analyzes Shimla’s unique congestion issues caused by steep terrain, narrow roads, and seasonal tourism. Using geostatistical tools like Kernel Density Estimation (KDE) and Moran’s I, the study found traffic volumes exceeding 395% of road capacity. The research advocates eco-friendly alternatives like ropeways, tunnels, and pedestrianized zones. Without such interventions, Shimla’s transit infrastructure risks collapse by 2030. The study underscores the need for integrated, multi-modal solutions in sensitive mountain ecosystems (Chandel, 2026).
Conclusion
Traffic congestion is not only caused by more vehicles but also by poor planning, rapid urban growth, and weak management. Studies show that building more roads is not a lasting solution, as it often attracts even more traffic. Instead, cities need smarter approaches like better public transport, eco-friendly travel options, and data-driven traffic systems.
Examples from Chennai, Shimla, and Peshawar highlight that without sustainable strategies, congestion will continue to harm productivity, health, and the environment. In simple terms, solving traffic problems requires a mix of technology, strong policies, and green infrastructure. If cities act early, they can make urban life smoother, healthier, and more efficient for everyone.
4.0 References
Chandel, A. S. (2026). Geostatistical analysis of traffic congestion in Shimla. Discover Sustainability, 7(97).
Chango, W., Bunay, P., Erazo, J., Aguilar, P., Sayago, J., Flores, A., & Silva, G. (2025). Predicting urban traffic congestion with VANET data. Computation, 13(4), 92.
Güven, G. (2025). Telework-induced urban sprawl and traffic congestion: A social discount rate analysis. European Transport Research Review, 17(44).
Huang, C.-J., Hu, K.-W., Ho, H. Y., Xie, B. Z., Feng, C.-C., & Chuang, H.-W. (2021). Mixed flow traffic congestion prevention mechanism. International Journal of Computational Intelligence Systems, 14(1).
Hussain, D., Jamal, A., Farooq, A., Almoshaogeh, M., Alharbi, F., & Farooq, D. (2026). Entropy-TOPSIS congestion mitigation framework. Scientific Reports, 16, 5036.
Jiang, D., Zhao, W., Wang, Y., & Wan, B. (2024). Spatiotemporal hierarchical analysis for congestion optimization. ISPRS International Journal of Geo-Information, 13(2), 59.
Parkavi, S., & Parthiban, A. (2025). Assessing urban traffic congestion for sustainable transportation in Chennai, India. Frontiers in Sustainable Cities, 7, Article 1684489.
Rossi, E., & Bianchi, M. (2025). Sustainable mobility approach to congestion hotspots. Future Transportation, 5(4), 138
Tsalikidis, N., Mystakidis, A., Koukaras, P., Ivanauskas, M., Morkūnaitė, L., Ioannidis, D., Fokaides, P. A., Tjortjis, C., & Tzovaras, D. (2024). Urban traffic congestion prediction using sensor data. Smart Cities, 7(1), 10.
Xu, Y., & Li, Y. (2025). A real-time urban traffic congestion prediction framework. IEEE Access, 13, 3608954.