Development and Evaluation of Vehicle to Pedestrian (V2P) Safety Interventions

Pedestrian deaths are on the rise with 6,227 estimated for 2018, the highest since 1990. Distractions such as walking while looking at electronic devices are the third leading cause of fatalities and recent research has shown that injuries from distracted walking have increased 81% since 2005. The introduction of self-driving cars could further complicate this problem as illustrated by the death of a pedestrian caused by an Uber self-driving car in 2018. To examine how well an electronic alerting device installed on a smartphone could prevent distracted pedestrians from making unsafe or risky crossings, an experiment was conducted in an actual controlled field setting. Using a smartphone with a remotely-controlled alerting system, thirty participants performed thirty crossings each while walking and playing a game on the smartphone. In addition to just-in-time alerts, two-thirds of participants were presented with early and late alerts which constituted 80% and 90% alarm reliabilities. Out of 900 crossing events, 20% of crossings were risky or unsafe. More than 18% of participants exhibited underestimation bias and thought the car was farther away than it really was. While international participants (i.e., non-US-born) as a group were more likely to attempt risky crossings while engaged in distracted walking, they also trusted the alert less when it generated early and late warnings. These results suggest that national origin may play an important role in the use of technological interventions meant to promote positive behaviors and that a solution effective in one setting may not generalize to other nations. Moreover, technology interventions like smartphone-based alerts do not produce substantially safer pedestrian behaviors than those observed in populations without such tools. While the subject pool was small in this study, this research suggests that there may be design criteria that can be elucidated from such use of machine learning classification methods in concert with controlled experiments.

Language

  • English

Media Info

  • Media Type: Digital/other
  • Edition: Final Report
  • Features: Figures; Photos; References; Tables;
  • Pagination: 18p

Subject/Index Terms

Filing Info

  • Accession Number: 01759938
  • Record Type: Publication
  • Report/Paper Numbers: CSCRS-R7
  • Contract Numbers: 69A3551747113
  • Files: UTC, TRIS, ATRI, USDOT
  • Created Date: Nov 23 2020 9:34AM