Clara Fang, PhD, Professor of Civil Engineering at the College of Engineering, Technology, and Architecture (CETA), and Daniel Jimenez Gil '24, an undergraduate senior in the civil engineering program, recently presented their research titled A Framework of Deep Neural Networks to Predict Bridge Condition Rating: Challenges and Opportunities at the 103rd Transportation Research Board (TRB) meeting held in Washington, D.C. The co-presenter also included Jacob Booth, a supervising transportation engineer in bridge management at the Connecticut Department of Transportation.
The research addresses challenges associated with predicting bridge condition ratings, including dealing with heterogeneous data sources, handling complex high-dimensional data, incorporating mixed types of data features, selecting features, and improving long-term prediction capabilities. The paper covers various aspects, including data analysis and organization, the proposed framework, the design and parameter tuning of the neural networks, and the deployment of the trained model. While the challenges and opportunities faced in this study are general to the problem, the authors will continue to focus on training the complete framework of neural networks and further deploying the trained model in the state of Connecticut.
The TRB Annual Meeting attracts thousands of transportation professionals from around the world, featuring more than 600 workshops, lectern sessions, committee sessions, and poster sessions, along with 200 exhibitions. The program covers all transportation modes, addressing topics of interest to policymakers, administrators, practitioners, researchers, and representatives of government, industry, and academic institutions. The US Secretary of Transportation, Pete Buttigieg, also attended this year's meeting and delivered an inspirational closing remark at the Chair's plenary session.