Scholarlink

<p>External human–machine interfaces (eHMIs) of highly automated vehicles (HAVs) have been developed to communicate vehicle intentions to pedestrians and other road users. This study conducted an online experiment with 33 participants to investigate the impact of a novel HAV’s eHMI of augmented reality (AR) crosswalks on pedestrian crossings using virtual simulation scenario videos and questionnaires. To verify the effectiveness of the AR crosswalk, comparative analyses were performed on eHMIs across 5 modalities: baseline, pedestrian silhouette, virtual eyes, AR headlight, and AR crosswalk, in conjunction with 2 vehicle physical modes (yielding and non-yielding). Metrics used in the study included crossing percentages before the HAV passed, crossing decision time, comprehensibility, and perceived safety. The results indicated that for yielding vehicles, the AR crosswalk resulted in the shortest crossing decision time, the highest comprehensibility, and the highest perceived safety scores. This study provides guidelines for the design of eHMIs and demonstrates that AR eHMIs based on an intelligent vehicle infrastructure cooperative system, can effectively address the one-to-many interaction problem.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Investigating the Effect of a Novel External Human–Machine Interface of Highly Automated Vehicles on Pedestrian Crossing: Augmented Reality Crosswalk

  • 期刊论文

  • 发表日期：

    2025年5月29日

查看全文
  • Hongyu Hu,
  • Xiaojie Diao,
  • Chuan Hu,
  • Luyao Wang,
  • Fei Gao,
  • Zhenhai Gao

摘要

External human–machine interfaces (eHMIs) of highly automated vehicles (HAVs) have been developed to communicate vehicle intentions to pedestrians and other road users. This study conducted an online experiment with 33 participants to investigate the impact of a novel HAV’s eHMI of augmented reality (AR) crosswalks on pedestrian crossings using virtual simulation scenario videos and questionnaires. To verify the effectiveness of the AR crosswalk, comparative analyses were performed on eHMIs across 5 modalities: baseline, pedestrian silhouette, virtual eyes, AR headlight, and AR crosswalk, in conjunction with 2 vehicle physical modes (yielding and non-yielding). Metrics used in the study included crossing percentages before the HAV passed, crossing decision time, comprehensibility, and perceived safety. The results indicated that for yielding vehicles, the AR crosswalk resulted in the shortest crossing decision time, the highest comprehensibility, and the highest perceived safety scores. This study provides guidelines for the design of eHMIs and demonstrates that AR eHMIs based on an intelligent vehicle infrastructure cooperative system, can effectively address the one-to-many interaction problem.