Project Information

Link to Latest Report: Coming Soon

Background:

A Generative Pre-trained Transformer (GPT) system can, if coupled with data, make a paradigm shift in our approaches to bridge design and construction. The research proposed in this project focuses on the development of BridgeGPT architecture, using bridge service life design as the first application. This is in keeping with US DOT Secretary of Transportation upper administration interest in the development of BridgeGPT, as expressed at the IBT/ABC-UTC kickoff meeting in May, 2023. At that meeting, the concept of BridgeGPT was included in the power point presentation.

The SHRP2 R19A project has gathered a wealth of data on bridge service life design. However, the full potential of this data has yet to be utilized. Traditional methods of bridge design and maintenance, while reliable, can benefit greatly from the predictive power and adaptability of models like BridgeGPT. There is a crucial need to more closely relate cuttingedge AI technology and bridge engineering to optimize service life design, reduce costs, and enhance infrastructure safety. This project aims to utilize the capabilities of GPT-like models, adapt them for bridge engineering, and create a tool that learns and improves over time.

Objective:

The research approach for this project focuses on developing the general architecture of the BridgeGPT platform, demonstrating the feasibility and usefulness of the BridgeGPT by developing a service life module compatible with the requirements of BridgeGPT. The data for service life design of bridges will be obtained from outputs of the SHRP2 R19A project. The beta version of the developed BridgeGPT platform and service life design module will be tested by select DOTs and consultants. The service life design module will be completed after receiving their feedback. This project will also include the development of a roadmap for future activities leading to the design of other modules, as well as a business plan for ensuring sustainability of the developed system.

Scope:

The proposed project includes several tasks

• Task 1: Developemnt of system architecture
  • This task will develop the overall architecture of a main platform that can accommodate
    different subsystems, each devoted to certain aspects of activities that are part of design
    and construction of bridges. Attempts will be made to develop an initial list of activities
    for which development of subsystem compatible with main platform requirements is
    feasible. Undoubtedly this list will expand in future; nevertheless, it will help to
    incorporate what is needed in the main platform to accommodate the potential
    subsystems.
• Task 2: Collection of service life design data
  • During the development of “Design Guide for Bridges for Service Life”, many interim and phase reports were prepared that were not included in the final report of SHRP2 R19A. These interim and phase reports contain a tremendous amount of useful data and information. Having served as PI of the SHRP2 R19A project, Dr. Azizinamini has access to complete documentations and has kept a good inventory of them. All data and information developed during conduct of SHRP2 R19A project will be reviewed, and available information will be categorized.
• Task 3: Development of general strategy for using AL and ML for service life design
  • To demonstrate the usefulness of the overall system, service life design of bridges will be used as the testbed. Under this task the strategy for how best to utilize principles of AI and ML for service life design of bridges in conjunction with available data will be developed.
• Task 4: Sharing beta version of service life design module with select users.
  • This task will share the interface to use service life design module with select DOTs and
    consultants who are willing to test the system and provide feedback.
• Task 5: Making changes and finilizing the service life design module
  • The feedback received will be used to modify and finalize the service life design module.
    This step may require several iterations.
• Task 6 : Developing roadmap for future work
  • This task will develop the roadmap for future activities, which will also include the
    business aspect of the initiative.
• Task 7 : Final Report
  • This task will summarize overall aspects of the project.

Research Team:
Principal Investigator: Atorod Azizinamini Ph.D., P.E.
Co-Principal Investigator: Mark Finlayson, Ph.D.