A case study of nighttime pedestrian automatic emergency braking performance under different roadway lighting and pedestrian clothing conditions

November 2024

Objectives: Automatic emergency braking systems with pedestrian detection (PAEB) reduce the rate of police- reported pedestrian crashes during the day but not at night (Cicchino 2022). This study evaluated how increasing pedestrian conspicuity using clothing or by increasing roadway lighting affected PAEB performance. Methods: The PAEB system in a 2023 Mazda CX-5, 2023 Honda CR-V, and a 2023 Subaru Forester were evaluated when an adult-sized mannequin crossed the road as each vehicle approached at 40 km/h. The mannequin was dressed in black, black with a retroreflective jacket, black with retroreflective strips in a biological motion configuration, or white. Light towers provided 0, 10, or 20 lux of average illumination in the crosswalk. The vehicle low beams were used in every combination of clothing and roadway lighting and high beams were used when the mannequin was in black without roadway lighting. The percentage of speed reduction was computed to compare PAEB performance between vehicles. Results: A collision occurred in 84% of trials with the CR-V, 88% of trials with the CX-5, and 2% of trials with the Forester. The Honda PAEB system reduced speed by 40% when the pedestrian was in black and high beams were used but did not reduce speed with low beams without roadway lighting. The system reduced speed somewhat with 10 and 20 lux of roadway lighting with the pedestrian in black or white clothing. The CX-5 PAEB system reduced speed by 68% when the pedestrian was wearing black and high beams were used but only by 30% with low beams without any roadway lighting. The CX-5’s PAEB system performance improved with increased roadway lighting when the pedestrian was wearing black. The CR-V and CX-5 PAEB systems did not reduce speed when the pedestrian was wearing retroreflective strips. The Subaru PAEB system reduced speed by 100% in all but one trial. Conclusions: Additional vehicle lighting enhanced PAEB system performance, but PAEB system performance with additional roadway lighting or more conspicuous clothing was inconsistent. Existing PAEB system hardware may be insufficient, or detection algorithms too brittle to cope with variations in the appearance of pedestrians at night.