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
Uncovering the Root Causes of Truck Rollover Crashes on Ramps
-
2023-03-01
-
[PDF-16.26 MB]
Details:
-
Creators:
-
Corporate Creators:
-
Contributors:
-
Corporate Contributors:
-
Subject/TRT Terms:
-
Publication/ Report Number:
-
Resource Type:
-
Geographical Coverage:
-
Edition:Final Report - March 2023 (03/01/2021-03/31/2023)
-
Corporate Publisher:
-
Abstract:Highway ramps are hotspots of truck rollover crashes. Such crashes often block the entire ramp and cause severe congestion. Understanding the major causes of ramp truck rollovers is critical for developing effective countermeasures. This research utilizes drones to collect ramp traffic videos at seven high-risk ramps and develops an Oriented Mask-RCNN model for vehicle detection and several other algorithms for tracking vehicles, extracting vehicle trajectories, and identifying high-risk events such as unsafe and last-minute lane changes. A thorough review of literature and best practices on reducing ramp truck rollovers is conducted. The narratives of all ramp truck rollovers between January 2015 and February 2022 in Massachusetts are reviewed. The locations and types of traffic signs, slopes, and curve radii of seven identified highway ramps are also obtained and carefully investigated together with the trajectory analysis results. It is found that over 95% of ramp truck rollovers are single-vehicle crashes and speeding is the predominant cause. Based on the analysis, some practical and specific safety improvement recommendations are provided. Drones are proven to be a useful and convenient tool for short-term traffic data collection. The developed vehicle detection and tracking algorithms work well to generate vehicle trajectories and detect high-risk events. The proposed suite of algorithms can also be applied to trajectories generated by other sensors such as radar. Overall, the proposed trajectory-based safety analysis approach provides important inputs for understanding driver behavior at/near highway ramps, and it can be used as an effective tool for road safety audits.
-
Format:
-
Funding:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type:
