A simulator-based analysis of engineering treatments for right-hook bicycle crashes at signalized intersections

• Published in: Accident analysis and prevention 2017-07, Vol.104, p.46-57
• Main Authors: Warner, Jennifer, Hurwitz, David S., Monsere, Christopher M., Fleskes, Kayla
• Format: Article
• Language: English
• Subjects: Accidents, Traffic - prevention & control; Adult; Aged; Attention; Automobile Driving - legislation & jurisprudence; Automobile Driving - psychology; Bicycle safety; Bicycling - psychology; Chi-Square Distribution; Computer Simulation; Driver behavior; Driving simulator; Engineering treatment; Environment Design; Female; Humans; Intersection design; Male; Middle Aged; Qualitative Research; Right-hook crash; Safety; Young Adult
• ISSN: 0001-4575; 1879-2057
• DOI: 10.1016/j.aap.2017.04.021

•The symbol sign increased side mirror scanning for the bicyclist by 9%.•White dotted lane markings outperformed white and green dotted for bike detection.•4% decrease in mean car velocity for moderate/high-risk incidents with 10ft radii.•There was a 54% smaller range of vehicle velocities for 10ft radii.•Protected intersections did not change the pattern of right-hook crash occurrence. A right-hook crash is a crash between a right-turning motor vehicle and an adjacent through-moving bicycle. At signalized intersections, these crashes can occur during any portion of the green interval when conflicting bicycles and vehicles are moving concurrently. The objective of this research was to evaluate the effectiveness of four types of engineering countermeasures – regulatory signage, intersection pavement marking, smaller curb radius, and protected intersection design – at modifying driver behaviors that are known contributing factors in these crashes. This research focused on right-hook crashes that occur during the latter stage of the circular green indication at signalized intersections with a shared right-turn and through lane. Changes in driver performance in response to treatments were measured in a high-fidelity driving simulator. Twenty-eight participants each completed 22 right-turn maneuvers. A partially counterbalanced experimental design exposed drivers to critical scenarios, which had been determined in a previous experiment. For each turn, driver performance measures, including visual attention, crash avoidance, and potential crash severity, were collected. A total of 75 incidents (47 near-collisions and 28 collisions) were observed during the 616 right turns. All treatments had some positive effect on measured driver performance with respect to the right-turn vehicle conflicts. Further work is required to map the magnitude of these changes in driver performance to crash-based outcomes.