Investigating the effect of urgency and modality of pedestrian alert warnings on driver acceptance and performance. (January 2019)

• Record Type: Journal Article
• Authors: Large, David R.; Kim, Hyungil; Merenda, Coleman; Leong, Samantha; Harvey, Catherine; Burnett, Gary; Gabbard, Joseph
• Is Part Of: Transportation research. Volume 60(2019)
• Page Start: 11
• Page End: 24
• Publication Date: 2019-01
• Subjects: Urgency -- Modality -- Driver acceptance -- Pedestrian alert systems -- HUD -- Vulnerable road users -- Simulation; Automobile drivers -- Psychology -- Periodicals; Automobile driving -- Psychological aspects -- Periodicals; Transportation -- Psychological aspects -- Periodicals; 629.283019
• DOI: 10.1016/j.trf.2018.09.028
• Languages: English
• ISSNs: 1369-8478

Abstract

Highlights: Earlier warnings increased the headway at which drivers responded. Greater confidence in combined audio and visual warnings compared to audio only. Lowest acceptance associated with warnings at shortest time-to-collision. No objective performance benefits associated with multi-modal warnings. Drivers accept high false positive alarm rates associated with early warnings. Abstract: Active safety systems have the potential to reduce the risk to pedestrians by warning the driver and/or taking evasive action to reduce the effects of or avoid a collision. However, current systems are limited in the range of scenarios they can address using primary control interventions, and this arguably places more emphasis in some situations on warning the driver so that they can take appropriate action in response to pedestrian hazards. In a counterbalanced experimental design, we varied urgency (' when' ) based on the time-to-collision (TTC) at which the warning was presented (with associated false-positive alarms, but no false negatives, or 'misses'), and modality (' how' ) by presenting warnings using audio-only and audio combined with visual alerts presented on a HUD. Results from 24 experienced drivers, who negotiated an urban scenario during twelve 6.0-min drives in a medium-fidelity driving simulator, showed that all warnings were generally rated 'positively' (using recognised subjective 'acceptance' scales), although acceptance was lower when warnings were delivered at the shortest (2.0 s) TTC. In addition, drivers indicated higher confidence in combined audio and visual warnings in all situations. Performance (based on safety margins associated with critical events) varied significantly between warning onset times, with drivers first fixating their gaze on the hazard, taking their foot off the accelerator, applying their foot on the brake, and ultimately bringing the car to a stop further from the pedestrian when warnings were presented at the longest (5.0 s) TTC. In addition, drivers applied the brake further from the pedestrian when combined audio and HUD warnings were provided (compared to audio-only), but only at 5.0 s TTC. Overall, the study indicates a greater margin of safety associated with the provision of earlier warnings, with no apparent detriment to acceptance, despite relatively high false alarm rates at longer TTCs. Also, that drivers feel more confident with a warning system present, especially when it incorporates auditory and visual elements, even though the visual cue does not necessarily improve hazard localisation or driving performance beyond the advantages offered by auditory alerts alone. Findings are discussed in the context of the design, evaluation and acceptance of active safety systems.