black and brown leather padded tub sofa

Realize intelligent interaction between vehicles and transportation infrastructure.

Optimize vehicle communication and infrastructure for safer roads.

A silver van is parked in an underground parking area, partially illuminated by overhead lights. The background reveals a multi-level parking structure with other vehicles visible through metal framework and lighting. Orange traffic cones and barriers are positioned around the entry control area.
A silver van is parked in an underground parking area, partially illuminated by overhead lights. The background reveals a multi-level parking structure with other vehicles visible through metal framework and lighting. Orange traffic cones and barriers are positioned around the entry control area.

Data Collection

Gather a comprehensive dataset of vehicle telemetry, traffic signals, road conditions, and environmental factors (e.g., weather, accidents) from urban and highway scenarios.

Vehicles are driving through an urban tunnel with overhead lights casting reflections on the road surface. The environment appears dimly lit, emphasizing the contrast between the vehicles' headlights and the dark surroundings.
Vehicles are driving through an urban tunnel with overhead lights casting reflections on the road surface. The environment appears dimly lit, emphasizing the contrast between the vehicles' headlights and the dark surroundings.

Model Fine-Tuning

Fine-tune GPT-4 on the V2I dataset to optimize its ability to analyze dynamic data, predict traffic behavior, and generate actionable insights for vehicles and infrastructure.

A modern black car with the label 'IONIQ 6' is parked in an industrial setting. Surrounding the vehicle are metal structures, storage silos, and piles of gravel. The area is open and quiet, with a mix of natural and industrial elements.
A modern black car with the label 'IONIQ 6' is parked in an industrial setting. Surrounding the vehicle are metal structures, storage silos, and piles of gravel. The area is open and quiet, with a mix of natural and industrial elements.

System Development

Develop an AI-powered V2I system that integrates the fine-tuned model to facilitate real-time communication and decision-making between vehicles and traffic infrastructure.

A street scene features a stationary white vehicle partially visible behind a metal railing. Nearby, another vehicle is parked under an overpass. Behind the railing, a blue and black building facade displays an electrical power sign. The surrounding environment includes some greenery and hints of urban infrastructure.
A street scene features a stationary white vehicle partially visible behind a metal railing. Nearby, another vehicle is parked under an overpass. Behind the railing, a blue and black building facade displays an electrical power sign. The surrounding environment includes some greenery and hints of urban infrastructure.

Expected Outcomes

This research aims to demonstrate that fine-tuning GPT-4 can significantly enhance the intelligence and efficiency of V2I systems. The outcomes will contribute to a deeper understanding of how advanced AI models can be adapted for real-time traffic management and vehicle-infrastructure interactions. Additionally, the study will highlight the societal impact of AI in improving road safety, reducing traffic congestion, and advancing smart city initiatives.

Insights

Optimizing traffic behavior through data analysis.

© 2025. All rights reserved.