Project ID: CTEDD 019-13
Author(s): Zhenyu Wang, University of South Florida
Co-Author(s): Pei-Sung Lin, University of South Florida; Srinivas Katkoori, University of South Florida
CTEDD Funding Year: 2019 General RFP
Project Status: Complete
UTC Funding: $109,743
End Date: May 31, 2020
Roadway lighting is a basic roadway infrastructure to ensure nighttime safety and security for all road users (motorists, pedestrians, cyclists, and transit passengers). To cost-effectively maintain a roadway lighting system, key tasks in infrastructure management include periodically measuring roadway lighting levels, diagnosing lighting performance based on collected data, and providing decision-making support for maintenance and improvement.
This project aims to develop innovative methods and tools to effectively and precisely recognize poor lighting patterns and predict associated nighttime crash risks using machine learning and deep learning models. Big lighting data collected for around 400 center-miles in Florida since 2012 (millions of measure points) will be used for core model training. Computer tools will be developed to integrate diagnosis models and data visualization functions. The tools will be tested in a real roadway environment and are expected to reach Technology Readiness Level 7: Prototype Demonstrated in Operational Environment.
This project will support multi-disciplinary collaboration for both faculty and students, including transportation engineering, computer science, and electrical engineering. The collaboration will also involve two stakeholders, including a government agency (Florida Department of Transportation [FDOT] District 7) and a private sector firm (Johnson, Mirmiran & Thompson, Inc. [JMT]), which will provide support and assistance for data collection, system design, diagnosis evaluation, and technology transfer.
A comprehensive technology transfer plan will be developed after the research, including implementation of the developed methods and tools in lighting measurement projects collaborated with stakeholders, presentations and publications, open source codes provided to the public, integration of research materials into coursework, and workshops in the Florida Local Technical Assistance Program (LTAP).
Through technology transfer, the project results will be beneficial to roadway lighting and safety managers for infrastructure performance monitoring and maintenance. The open algorithms and source codes also will be beneficial to researchers and practitioners for future research and practice.