United States Department of Transportation
ROSA P serves as an archival repository of USDOT-published products including scientific findings, journal articles, guidelines, recommendations, or other information authored or co-authored by USDOT or funded partners.
As a repository, ROSA P retains documents in their original published format to ensure public access to scientific information.
i
Snow Plow Performance Measures in Non-RWIS Locations
-
2024-08-01
-
[PDF-2.18 MB]
Details:
-
Creators:
-
Corporate Creators:
-
Corporate Contributors:
-
Subject/TRT Terms:
-
Publication/ Report Number:
-
Resource Type:
-
Geographical Coverage:
-
Edition:Final Report September 2023 to August 2024
-
Corporate Publisher:
-
Abstract:This project aims to enhance traffic safety by developing an Artificial Intelligence (AI) model and the weather interpolation method to evaluate snow cover conditions on road surfaces using existing roadside Closed-Circuit Television system (CCTV) images in non-Road Weather Information System (RWIS) locations. Snow- cover significantly impacts traffic safety, contributing to 24 percent of annual weather-related vehicle crashes. In Utah, where snow seasons can last up to seven months with over 25 winter storms annually, understanding snow- cover condition in real time is crucial for effective snowplow management by the Utah Department of Transportation (UDOT) and for public safety. Traditional methods for assessing road conditions rely on RWIS, which has limited coverage. The project will utilize image data from CCTV cameras to train a deep learning-based AI model for automatic evaluation of snow-cover condition. The methodology includes processing images to create labeled datasets, developing AI models based on AlexNet as the deep learning algorithms, and implementing these models to analyze real-time snow-cover conditions. Additionally, a weather interpolation method is developed to estimate real-time weather status in non-RWIS areas. The validation results show that our developed AI model can achieve over 97 percent accuracy in identifying various snow-cover states, including clear, partial snow, and full snow conditions. Furthermore, the developed weather interpolation method can provide a detailed view of the spatial distribution of air temperature, relative humidity, wind speed, precipitation, and snow accumulation. By leveraging existing CCTV networks, this project offers a cost-effective solution for large-scale, real-time road condition monitoring, closing information gaps, and enhancing winter road safety management.
-
Format:
-
Funding:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type:
