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
Dynamic Travel Time Estimation for Northeast Illinois Expressways
-
2020-06-01
[PDF-3.47 MB]
Details:
-
Creators:
-
Corporate Creators:
-
Corporate Contributors:
-
Subject/TRT Terms:
-
Publication/ Report Number:
-
DOI:
-
Resource Type:
-
Geographical Coverage:
-
Edition:Final Report 9/1/17–6/30/20
-
Corporate Publisher:
-
Abstract:Having access to accurate travel time is critical for both highway network users and traffic operators. Travel time that is currently reported for most highways is estimated by employing naïve methods that use limited sources of data. This might result in inaccurate travel time prediction and could impose difficulties on travelers. The purpose of this report is to develop an enhanced travel time prediction model using multiple data sources, including loop detectors, probe vehicles, weather condition, geometry, roadway incidents, roadwork, special events, and sun glare. Different models are trained accordingly based on machine learning techniques to predict travel time 5 min, 10 min, and even 60 min ahead. A comparison of techniques showed that 15 min or shorter prediction horizons are more accurate when applying the random forest model, although the prediction accuracy of longer prediction horizons is still acceptable. An algorithm is proposed for dynamic prediction of travel time in which the travel time of each highway corridor is calculated by adding the predicted travel time of each link of the corridor. The proposed dynamic approach is tested and evaluated on highways and showed a significant improvement in the accuracy of predicted travel time in comparison to the snapshot travel time prediction approach. Traffic-related variables, especially occupancy, are found to be effective in short-term travel time prediction using loop-detector data. This suggests that among traffic variables collected by loop detectors, occupancy can capture traffic condition better than other variables. Fusion of several data sources, however, increases prediction accuracy of the models.
-
Format:
-
Funding:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type:
