Objectives: The aim of the study was to establish whether the driver's visual strategy may influence a driver's behavior to avoid a crash in a high-risk situation. Any published papers on drivers' visual strategies just before a crash were not found.
Material and methods: Tests were performed using a high-tech driving bus simulator. Participants comprised 45 men drivers, aged 43.5±7.9 years old, seniority as a bus driver of 13.3±8.6 years. The tests were preceded by medical examinations: general, neurological and ophthalmological. Each participant drove the same city route for approximately 40 min (entire route - ER). In the final phase, a collision situation was simulated (a phantom car blocked the participant's right of way). Driver's visual strategy was analyzed using the FaceLab device with 2 cameras during ER and just before collision. The field-of-view covered by camera 1 was divided into 8 regions, by camera 2 into 10 regions. The distribution of gazes in regions was a criterion of visual strategy.
Results: Thirty-five drivers completed the simulated driving test, 14 escaped the collision, 21 crashed. These groups differed only in resting systolic blood pressure before the test. The analysis of covariance, after adjusting to this factor, indicated that during the ER visual strategy recorded by camera 1 did not differ between groups, in camera 2 the drivers in the crash group fixed their gaze more frequently (p = 0.049) in region 3 (close part of the road in front of the windshield). Just before the collision drivers who escaped the collision fixed their gaze significantly more often in region 6 (left side of the road) in camera 1 and in region 6 (in front of the windshield,) and region 10 (right side) in camera 2.
Conclusions: The visual strategy has an impact on the road safety. The analysis of visual strategies may be a useful tool for the training of drivers. Int J Occup Med Environ Health. 2019;32(2):161-74.
Keywords: accidents; driver; driving simulator; eye-tracking; road safety; visual strategy.
This work is available in Open Access model and licensed under a CC BY-NC 3.0 PL license.