Abstract
Purpose :
People with central vision loss (CVL) are permitted to drive in most states in the USA; however, CVL may cause delayed responses to hazards. Advanced driver assistance systems that provide hazard warnings are available, but little is known about the extent to which such systems might mitigate crash risk of drivers with CVL. We therefore developed a prototype hazard warning device and evaluated its impact on the responses of drivers with CVL to pedestrian hazards in a simulator.
Methods :
Seven subjects with CVL (36 – 74 y; median VA 20/80) and 7 subjects with normal vision (NV) (34 – 78 y; median VA 20/15) participated. They completed 2 drives with and 2 drives without a custom-designed vibrotactile warning device embedded in the seat cushion. There were 10 realistic pedestrian hazards per drive which appeared mid-block from either the left or right and crossed the road in front of the driver’s vehicle, requiring a natural braking response by the driver to avoid a collision. If collision risk (computed continuously) exceeded a pre-defined threshold, the device gave a directional vibration to warn of the approaching hazard. Gaze position and driving metrics were monitored at 50 Hz for analysis. The virtual city environment was populated with other distractor pedestrians, crowds, and traffic.
Results :
Overall, CVL subjects took longer to fixate hazards (1.2s vs 0.6s, p = .01), received more warnings (77% vs 66%) and were involved in more collisions (2.9% vs 0.4%, p = .01) than NV subjects. When collision risk exceeded the threshold, warnings had no effect on the time to fixate the hazard, but significantly reduced the time taken by CVL subjects to press the brake (from 2.4s to 2.0s, p = .004). Thus, the warnings significantly improved response safety of CVL subjects (p < .001), reducing their collision rate from 8.5% to 0% for events exceeding the collision threshold. In contrast, the warnings had no effect on safety of NV drivers (interaction p < .001). CVL subjects liked the design of the warnings (intensity, duration, timing and frequency), rated the device as more helpful than NV subjects, and would be more likely to use it if available on their car.
Conclusions :
Our results provide preliminary evidence that tactile directional warnings of potential hazards may reduce collision risk of CVL drivers by decreasing the time required to perceive the risk level and initiate a braking response.
This is a 2021 ARVO Annual Meeting abstract.