INVESTIGATION OF NEAR-CRASH TRAFFIC EVENTS BASED ON NATURALISTIC DRIVING DATA

Simple item page
dc.contributor.advisorD. Prevedouros , Panos
dc.contributor.authorCarneiro Pereira, Luana
dc.contributor.departmentCivil Engineering
dc.date.accessioned2021-09-30T18:02:05Z
dc.date.issued2021
dc.description.degreeM.S.
dc.embargo.liftdate2022-09-29
dc.identifier.urihttp://hdl.handle.net/10125/76276
dc.subjectCivil engineering
dc.subjectTransportation
dc.subjectdistraction
dc.subjectNaturalistic driving studies
dc.subjectnear-crash
dc.subjectrear-end
dc.subjectregression
dc.titleINVESTIGATION OF NEAR-CRASH TRAFFIC EVENTS BASED ON NATURALISTIC DRIVING DATA
dc.typeThesis
dcterms.abstractThis study collected naturalistic driving data derived from the collaboration of University of Hawaii of Manoa (UHM) and Charley’s Taxi and Limousine (CTL). Dashboard cameras and sensors were installed in 233 vehicles in Oahu, Hawaii and several hours of naturalistic driving data (NDD) were recorded in a period of seven months. The main goals of this thesis were to develop a database based on the naturalistic driving data, observe the main near-crash events recorded, and identify the main factors contributing to the near-crash/crash occurrence. Several studies done previously in different part of the world have shown that through NDD, it is possible to analyze near-crashes event to understand crash risky factors, which is important as the number of crashes recorded data is low. This study developed a database with a total of 402 harsh events, 398 near-crashes and 4 crashes. Several variables such as road characteristics, environmental characteristics, driver characteristics and vehicle characteristics were coded for each event. The following objectives were set to achieve the study purpose: (1) collect data from NDD events where driving maneuvers caused an acceleration or deceleration of 0.5g or higher; (2) develop a database suitable for research analysis; (3) derive basic statistics for all variables; (4) investigate correlations between variables; and (5) further investigate the variables related to the most frequent types of events using stepwise linear regression models. The main findings of this study may be summarized as follows: • Nearly 18% of events occurred on Thursdays and only 12% occurred on Sundays. • Only 17.2% of the events occurred on a curvy road segment, while 82.8% occurred on a straight road segment. • About 7.5% of the total events occurred on a construction zone or a blocked lane. • About 31% of the events occurred on a road segment with no traffic control and about 16% occurred on a freeway. • Signal traffic control was present in about 40% of the events. • About 47% of the events occurred on roads with medium traffic congestion, and 29% under heavy traffic congestion. • Pavement surface was wet in 10% of the events. • Lighting conditions were as follows: 13% cloudy, 16% dark, 63% sunlight/clear, 6% glare/sunset/sunrise, and 1% tunnel. • Pedestrians were present in 24% of the events. • The auto braking did not seem to have a significant impact in event occurrence. • We tested weather there was a correlation between the human perception of severity represented by the variable level of perceived impact and the machine severity level of severity represented by the G-force recorded by the sensors installed in the vehicles. A positive correlation with R2 of about 0.4 was found for near rear-end events on freeways. • The most frequent type of event was V1 almost rear-ended V2, followed by lane changing related events, and events with pedestrians. • Near rear-end events on freeways are common. The most influencing factors are light traffic conditions (negative effect) and mobile phone use (positive effect). • Lane changing events on freeways are strongly affected by mobile phone use and absence of automated driver aids in the vehicle. Keywords: Naturalistic driving studies, near-crash, rear-end, regression, distraction.
dcterms.extent76 pages
dcterms.languageen
dcterms.publisherUniversity of Hawai'i at Manoa
dcterms.rightsAll UHM dissertations and theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission from the copyright owner.
dcterms.typeText
local.identifier.alturihttp://dissertations.umi.com/hawii:11159

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
CarneiroPereira_hawii_0085O_11159.pdf
Size:
1.3 MB
Format:
Adobe Portable Document Format
Download

Collections

M.S. - Civil Engineering