Skip to main content
ICTD 2024 Main banner
Icon Legend
This session is not in your schedule.
This session is in your schedule. Click again to remove it.

Abstract for presentation (Poster or Podium)

AI in Transportation

Transforming Conversational Inputs into Dynamic Traffic Simulations Using Large Language Model

Monday, June 17, 2024
1:30 PM - 2:45 PM ET
Room: Galleria

    Primary Author(s)

  • Talha Azfar

    Ph.D. Student
    Rensselaer Polytechnic Institute
    Troy, NY, United States

    • Presenter(s)

  • Talha Azfar

    Ph.D. Student
    Rensselaer Polytechnic Institute
    Troy, NY, United States

    • Co-Author(s)

  • RK

    Ruimin Ke, n/a

    Assistant Professor
    Rensselaer Polytechnic Institute
    Troy, New York, United States

    • Submitter(s)

  • RK

    Ruimin Ke

    Rensselaer Polytechnic Institute
    Troy, New York, United States

Abstract: Large Language Models (LLMs), capable of handling multimodal inputs and outputs such as text, voice, images, and video, are transforming the way we process information. Beyond just generating textual responses to prompts, they can integrate with different software platforms to offer comprehensive solutions across diverse applications. Our methodology begins by creating a rich training dataset, leveraging the LLM for diverse inputs, and using Python scripts for outputs. These scripts are designed to convert specified regions into coordinates, fetch data from OpenStreetMap, transform it into a road network, and subsequently run the SUMO simulation with the designated traffic conditions. We then finetune Mistral 7B, an open-source LLM, on this curated dataset. This enables Mistral 7B to execute SUMO simulations for real-world locations based on conversational human inputs, retaining the context for continued interaction. The resulting visual simulation is presented, complemented by concise and intelligible summaries of the results.

Learning Objectives:

Attendees can expect to learn the following from this session:
  • Upon completion, participants will effectively utilize LLMs for multimodal inputs like text, voice, and visuals.
  • Upon completion, participants will execute Python scripts to convert data, utilize OpenStreetMap, and run SUMO simulations.
  • Upon completion, participants will fine-tune LLMs like Mistral 7B, enabling conversational input-driven real-world simulations.