TOWARDS IMPROVING PEDESTRIANS’ SAFETY: INSIGHTS ON PEDESTRIANS’ INTERACTIONS WITH AUTONOMOUS VEHICLES ON HIGH-SPEED ARTERIALS

Pedestrian safety on high-speed arterials (≥40 mph) is a growing concern due to the alarming rise in pedestrian-related crashes, with excessive vehicle speeds being a major contributing factor. While national guidelines offer countermeasures for low-speed roads, guidance for high-speed arterials is limited. To improve pedestrians’ safety on high-speed arterials, this study aims to contribute to literature by achieving two main objectives (1) to thoroughly examine the best practices and effective countermeasures that can be implemented to reduce pedestrian risk, (2) to examine whether pedestrian crossing behaviors differ in the presence of autonomous vehicles (AVs) compared to human-driven vehicles (HDVs). First, a nationwide survey among state Departments of Transportation (DOTs) professionals was conducted. Responses from 48 state DOTs revealed most states rely on the MUTCD for policy guidance. Driver errors, pedestrian errors, and lack of pedestrian facilities were identified as key crash factors. Key criteria for providing pedestrian facilities included pedestrian volume, traffic volume, and crash history. Fatality data from all U.S. states was collected to examine the association between pedestrian safety countermeasures and fatality rates. Stepwise multiple regression on FARS data from 2017-2022 revealed that narrowing road widths at midblock is significantly associated with reduced fatality rates. Providing medians and high-visibility crosswalks were linked to lower fatality rates in dark not-lighted midblock, while curb extensions and pedestrian countdown signals were effective at intersections, particularly in dark not-lighted conditions. In addition, a VR experiment was developed to investigate whether pedestrians’ crossing behaviors and trust differ in the presence of AVs compared to HDVs. Fifty-six participants completed six crossing scenarios. The results from the Generalized Linear Mixed Model showed significant differences in behavior towards AVs and HDVs. Pedestrians had longer post-encroachment times (PET) with AVs but shorter waiting times at the curb and median, leading to faster crossing times and fewer crashes with AVs. Pedestrians aged ≥25 took longer to cross, and males waited longer at the median. Low visibility increased caution and reduced risk, as shown by shorter PET. Post-experiment survey responses indicated higher pedestrian confidence in HDVs, but AVs scored higher in predictability, dependability, and reliability.