In an era where highways symbolize economic momentum and urban expansion, traffic congestion has quietly evolved into one of the most pressing infrastructural challenges of our time. It drains fuel, increases carbon emissions, heightens accident risk, and erodes productivity. For Vyom Shrivastava, a recent Engineering Graduate from Amity University, this was not merely a civil inconvenience but a complex systems problem demanding an intelligent, scalable solution. His project, AI Driven Traffic Prediction and Vehicle Speed Advisory, reflects the convergence of predictive analytics, computer vision, and systems engineering to create safer, greener highways.
His project exemplifies the Academy’s commitment to nurturing student innovators who combine technical rigor with measurable societal impact.
Vyom’s work is grounded in a powerful insight: traffic patterns are temporal and predictable when analyzed with the right intelligence frameworks. To capture these dynamics, he designed a two subsystem AI architecture that integrates predictive modeling with real- time visual analytics.
At the core of the predictive layer lies a Long Short Term Memory model, chosen for its strength in analyzing sequential time series data. The model was trained on the Metro Interstate I 94 dataset in the United States and rigorously validated across major Indian highways including NH 1, NH 19, NH 44, NH 48, and the Yamuna Expressway. Comparative benchmarking demonstrated that the LSTM architecture significantly outperformed CNN and ANN baselines on Mean Squared Error and Mean Absolute Error metrics, proving its ability to capture traffic flow patterns across diverse geographies and driving behaviors.
Yet forecasting alone does not translate into impact. To bridge prediction with reality, Vyom integrated a real time perception layer powered by YOLOv8 dashcam analytics. This module performs live vehicle detection and counting, enabling continuous recalibration of on road density. By combining predicted congestion levels with real time vehicle counts and roadway attributes, the system generates dynamic traffic coefficients that inform speed advisory decisions in the moment.
The most distinctive innovation lies in the engine aware speed mapping framework. Through controlled field experiments on NH 8 using instrumented vehicles across three engine types, Vyom developed traffic to optimal speed mappings grounded in fuel efficiency metrics such as Brake Specific Fuel Consumption. Instead of issuing uniform advisories, the system recommends vehicle specific speed envelopes aligned with engine efficiency zones. This approach enhances fuel economy while maintaining safety and travel efficiency.
Equally compelling is the model’s cross geography validation design. Training on high resolution United States data and validating across Indian highways demonstrated transferability and robustness. The system adapts to variations in infrastructure design, traffic culture, and density dynamics, reinforcing its scalability.
From an implementation perspective, the architecture remains lightweight and deployment ready. The LSTM forecasting engine operates on preprocessed time series inputs accounting for seasonal variations and missing data. YOLOv8 enables efficient edge inference directly from dashcam feeds. A compact advisory engine integrates traffic coefficients, lane parameters, and engine profiles to compute stable, lane specific speed bands. Field validation shows reductions in abrupt braking events and improved flow stability, contributing to smoother journeys and measurable fuel savings.
Through this work, Vyom Shrivastava demonstrates how artificial intelligence can move beyond theoretical modeling to real world optimization. His architecture lays a foundation for cooperative intelligent transport systems, vehicle to infrastructure communication, and fleet level fuel efficiency enhancement.
AIM Elevate Academy celebrates Vyom’s technical achievement and responsible innovation. His project embodies the Academy’s ethos: engineering solutions that are intelligent, impact driven, and aligned with sustainable development goals.
The highways of tomorrow will not merely carry traffic.
They will carry intelligence shaped by innovators like Vyom Shrivastava.