This study assessed the appropriateness of two warning timing algorithms and evaluated various interface approaches for presenting and displaying warning information to automobile drivers. Testing was conducted in a controlled environment using two instrumented vehicles, both equipped with an RPA (rear park assist) feature and a longer-range prototype backing warning system. Testing was conducted under both expected and surprise conditions. Under the expected conditions, drivers were asked to back to known obstacles (a child-sized mannequin and a simulated wall) and brake in response to the warning alerts. Results suggested that the predicted timing condition (in which drivers were provided with more advance warning time) led to more acceptable ratings, fewer target strikes and close calls, and less urgent braking. In the surprise event testing, a toy coupe was launched into the vehicle's path as drivers were backing up, causing the RODS (rear obstacle detection system) to issue a warning to the driver. Across the trials in which a warning was issued, only 13% of drivers avoided hitting the toy coupe (5 of 39); more than 87% of the drivers collided with the toy coupe following the warning. Although a large percentage of drivers braked (44%) in response to the warning, the level of braking was generally insufficient to avoid collision with the toy coupe.
Samenvatting