Purchase this article with an account.
Jing Xu, Vilte Baliutaviciute, Garrett Swan, Alex R Bowers; Effects of intersection cross traffic on scanning and responding to pedestrian hazards by drivers with hemianopia. Invest. Ophthalmol. Vis. Sci. 2020;61(7):914.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
In prior studies, drivers with hemianopic field loss (HFL) exhibited deficits in blind-side scanning resulting in impaired detection of hazards at intersections. Cross traffic is common in urban driving but the effects of cross traffic were not previously investigated. Using a driving simulator we investigated the effects of intersection cross traffic on scanning behaviors and responses to high-fidelity pedestrian hazards by drivers with HFL.
To date, 12 people with complete HFL and 7 drivers with normal vision (NV) have participated. They drove through 84 four-way intersections in a busy urban environment with critical events at 30 intersections while eye and head movements were tracked. Cross traffic approached from the left and/or right at 36 km/h. A pedestrian standing on the far side of the intersection, on the left or right, suddenly ran across the road as soon as the subject started driving through the intersection, requiring a braking response to avoid a potential collision. Three types of cross traffic were programmed: (1) intersections without cross traffic, (2) intersections with one approaching car from the side opposite the pedestrian location, and (3) intersections with two approaching cars, one from each side at the same time. To decrease anticipation of critical events, cross traffic was programmed at non-critical intersections, and there were distractor pedestrians and pedestrian crowds.
HFL drivers looked at the majority of cross traffic and pedestrian hazards on both the blind and seeing sides. Interestingly, when the HFL drivers fixated on both approaching cars, they responded more safely to the pedestrians (p < .001) and made more numerous and significantly larger gaze scans when compared to looking at one or no cars (p < .001 including data from all three types of cross traffic). Overall, HFL drivers made a greater number of scans to the blind side than NV drivers to the corresponding side (p = .005) and their scan magnitudes were larger (p < .001) but they were slower to make their first fixation on pedestrians (p = .035).
Cross traffic from both directions did not distract HFL drivers. Rather, it appeared to have beneficial effects in encouraging HFL drivers to make more and larger scans, and enter the intersection more cautiously with safer responses to pedestrians.
This is a 2020 ARVO Annual Meeting abstract.
This PDF is available to Subscribers Only