Research Papers

Visual Attention Allocation during Right Turns at Signal-Controlled Intersections: An On-Road Study

Filename 1B-Kaya-FP.pdf
Filesize 412 KB
Version 1
Date added July 9, 2018
Downloaded 6 times/fois
Category 2018 CARSP XXVIII Victoria
Tags Research and Evaluation, Session 1B
Author/Auteur Kaya, Ayas, Ponnambalam, Donmez
Stream/Volet Research and Evaluation

Slidedeck Presentation Unavailable


"Negotiating an urban intersection is a high demand maneuver. Toronto Public Health (2015) reported that between 2008 and 2012, 69% of motor vehicle collisions with vulnerable users occurred at intersections, and most included a turning vehicle. To understand where drivers allocate attention when performing a right turn at a signal-controlled intersection across a major arterial road and a side street; to what level drivers perform necessary checks for vulnerable users; and how traffic light status during approach and turn affects attention allocation. Eye-tracking data was analyzed from 17 drivers who participated in an instrumented vehicle study conducted in downtown Toronto. The focus was on glances during right-turns on a specific intersection (turn-on-red allowed, dedicated bike lane on the major arterial). Each driver made a right-turn on this intersection both to the major arterial (from the side street) and to the side street (from the major arterial). Glances were categorized as: windshield, left-side window and mirror, right-side window and mirror, rear-view mirror. During turns to the major arterial, driver attention was mostly divided between the left and the front (30% and 64% of total glance time on the average, respectively); whereas drivers looked mostly to the right (17%) and the front (76%) during turns to the side street. Drivers failed to check their rear-view mirror and right-hand side in more than half of the turns to the major arterial (10 out of 17); compared to 3 out of 17 for turns to the side street.
A breakdown by light status revealed that when making a right turn to the major arterial, drivers spent more time looking at the left when the light was red (35% of total glance time) compared to when it was green (6%); they also spent less time looking towards the front (61% vs. 91%). Light status had an effect on the rear-view mirror checks during turns to the side street (4% on green vs. 1% on red). The higher time spent looking towards the left during turns to the major arterial were expected due to the busier traffic. During turns to the side street, drivers spent more time looking towards their right as the major arterial had a dedicated bike lane. However, even for these turns, some drivers failed to do a cyclist check; the failure rate was much higher (~59%) when turning to the major arterial.
When the turn to the major arterial was on green, drivers tended to direct their attention towards the front to the pedestrians crossing the major arterial. When drivers were turning to the side street, they were more likely to check their rear-view mirror when the light was green compared to red, potentially watching for cyclists approaching or being concerned for blocking traffic. Drivers appeared to allocate their attention based on visual complexity and their expectations. However, necessary checks were not performed for vulnerable users in a considerable number of turns. Further research with an increased sample size and on a variety of intersections is needed to generalize these findings.