Abstract
Purpose :
Patients with visual field loss report difficulties in navigating open spaces, where they tend to collide with other pedestrians. A better understanding of the visual cues people use in detecting pedestrian collisions will benefit the design of field expansion devices. In a preliminary investigation, we measured the responses of normally-sighted subjects as they made pedestrian collision judgments.
Methods :
We configured an open-space virtual walking using a driving simulator. The simulated pedestrians initially appeared at the subject’s bearings of 10°, 30°, or 45°, and they were on collision courses with the subject (time-to-collision 6 s) or they aimed to pass the subject either behind or in front. The distance between the subject and pedestrian when it crossed the subject’s path (defined as the cross-path distance, or CPD) was tested at -4, 0, +4, and +12 m. Subjects (N = 5) indicated on each trial whether the simulated pedestrian would collide with them. Both the response time (RT) when the subjects detected the pedestrians and the time when they made collision judgments were recorded.
Results :
As for subjects’ perceived collision decisions, significant main effects of the CPDs, bearings, and their interactions were observed (all p < 0.001). More perceived collisions were reported in the conditions with a 10° initial bearing, which is consistent with the presence of a stronger looming effect in these conditions. Given a large initial bearing, the rate of perceived collision was larger when the pedestrians passed in front (CPD +4 m) than behind (-4 m), which may be due to slower bearing changes while passing in front. Across all trials, the RT for pedestrian detection was 1.0 ± 0.5 s, and the decisions were made 2.0 ± 1.6 s after the detection. The detection RT was similar across all conditions. However, the decision RT was significantly influenced by the bearings (p = 0.01), CPDs (p < 0.001), and their interactions (p = 0.003). A longer decision time was required when the pedestrians passed in front at +4 m CPD. Additionally, comparing with collision decisions, the non-collision decisions took slightly longer to make (p < 0.001).
Conclusions :
Normally-sighted observers take both looming and bearing cues into consideration when judging pedestrian collisions. Oncoming pedestrians are more likely perceived as colliding when the bearings change more slowly and/or a stronger looming appears.
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.