United States Department of Transportation
i
Deep Learning Software for Traffic State Prediction
-
2023-08-01
[PDF-8.34 MB]
Details:
-
Creators:
-
Corporate Creators:
-
Corporate Contributors:
-
Subject/TRT Terms:
-
Publication/ Report Number:
-
Resource Type:
-
Geographical Coverage:
-
Edition:2/1/2020 – 2/28/2023
-
Corporate Publisher:
-
Abstract:As urban populations continue to grow, traffic congestion has become an increasingly pressing issue for drivers, leading to a greater demand for advanced navigation aids and traffic management solutions. Real-time traffic information, advanced traffic signal control strategies, and the integration of various data sources can play a crucial role in improving traffic flow and safety. However, these advancements also present challenges, such as data integration and effective analysis. This research explored the development of deep neural network-based software that converts video into structured data formats, providing valuable information on driving conditions and traffic incidents. The use of synthetic data in response to the absence of normal traffic patterns during the COVID-19 pandemic is also examined. The research products of this project include a video analytics pipeline, robust data fusion techniques, and a prototype application designed to work with actual traffic control hardware, ultimately aiming to enhance traffic flow and safety on our roads.
-
Format:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type:
