Effective management and maintenance of traffic signs are essential for ensuring road safety and operational efficiency. However, traditional, manual methods for assessing traffic sign conditions are labor-intensive, time-consuming, expensive, often inconsistent, and they can pose safety risks for personnel. To address these challenges, the researchers have developed a novel Automatic Traffic Sign Detection and Assessment System (ATSDAS). The ATSDAS uses video logs and GPS data and leverages AI-driven edge computing, deep learning models, and feature extraction to identify, classify, and evaluate the condition of traffic signs in real time. It also organizes and stores traffic sign data dynamically, enabling effective asset management for transportation agencies. The focus of this project is to validate and refine the previously developed system for onboard traffic sign detection and condition assessment by using an edge device. This work will transition the system from a laboratory setting to practical field applications, emphasizing its robustness, reliability, and scalability for real-world asset management. By integrating this innovation into statewide operations, Washington state will not only improve road safety and operational efficiency but also position itself as a leader in leveraging AI-driven solutions for transportation infrastructure.
Principle Investigator: Yinhai Wang, Civil and Environmental Engineering, UW
Sponsors: WSDOT / FHWA STIC
WSDOT Technical Monitor: Trevor McCain
WSDOT Project Coordinator: David Strich
Scheduled completion: April 2027