Exploring the association of rear-end crash propensity and micro-scale driver behavior. (November 2016)

• Record Type: Journal Article
• Authors: Kim, SangKey; Song, Tai-Jin; Rouphail, Nagui M.; Aghdashi, Seyedbehzad; Amaro, Ana; Gonçalves, Gonçalo
• Is Part Of: Safety science. Volume 89(2016)
• Page Start: 45
• Page End: 54
• Publication Date: 2016-11
• Subjects: Rear end crashes -- In-vehicle sensing -- Driver behavior -- Safety; Industrial accidents -- Periodicals; Accident Prevention -- Periodicals; Safety -- Periodicals; Travail -- Accidents -- Périodiques; 363.11
• DOI: 10.1016/j.ssci.2016.05.016
• Languages: English
• ISSNs: 0925-7535

Abstract

Highlights: This paper provides initial evidence of an association between rear-end crash rates and driver behavior. 85% of all rear-end crashes occurred within 2000 feet of the on-ramp gore. There was a strong association between rear-end crash rates and the propensity of hard deceleration (below 4/m/s 2 ). The extreme deceleration problem appears to be consistent across most drivers traveling some of the links. Abstract: The relationship between driver behavior at the tactical level and crash experience is a long sought association that has been elusive to explore. The availability of in-vehicle sensing devices capable of capturing and documenting micro-scale dynamic driver behavior offers the opportunity to begin such an exploration. This study integrates rear-end crash data experienced on a 63-mile section of I-40 in North Carolina over a four-year period with three months of micro-scale driving behavioral data gathered by an in-vehicle sensing system (i2D) that records and dispatches second by second vehicle dynamics data to a central database. The information collected by the i2D devices came from a fleet of about 20 vehicles driven by volunteers in their naturalistic driving environment. Additionally all crash and driver data were geo-located onto a link-based GIS environment. The objective of this study is to explore the association of crash propensity and micro-scale driving behavior. The initial findings of this research are promising. First, over 85% of all rear-end crashes occurred on 30 segments extending from 2000 feet upstream of an on-ramp to the on-ramp itself. Secondly, on those segments with high crash rates we have detected a high propensity of drivers to decelerate at high rates (4 m/s 2 or more). We have also tested and confirmed that the sharp deceleration phenomenon is not confined to a few drivers, but appears to be common for the high-crash segments, using trip-based analyses.