Response of Vulnerable Road Users to Visual Information from Autonomous Vehicles in Shared Spaces

Alvarez, Walter Morales; de Miguel, Miguel Angel; Garcia, Fernando; Olaverri-Monreal, Cristina · 2019 · OpenAlex-citations

DOI: 10.1109/itsc.2019.8917501

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study investigates the impact of visual communication cues from autonomous vehicles (AVs) on the behavior of vulnerable road users (VRUs), specifically pedestrians, in shared spaces lacking traditional traffic infrastructure. The research addresses the challenge of establishing trust and safety in environments where pedestrians cannot make eye contact with a driver. The authors hypothesized that pedestrians would be more likely to pause or yield to an AV if the vehicle did not signal that it had detected them. To test this, the study examined whether specific visual displays—such as traffic-light colors or open/closed eyes—significantly altered pedestrian crossing behavior compared to a baseline condition with no display. The experimental design involved field tests conducted over two days at the University Carlos III campus in Madrid, a shared space where informal traffic rules prevail. The researchers used the Intelligent Campus Automobile (iCab), an autonomous vehicle equipped with an external Human-Machine Interface (HMI). The HMI displayed three conditions: a baseline (no message), a red-green sign indicating safety, and open-closed eyes mimicking driver detection. The vehicle traveled a 30-meter route at speeds not exceeding 5 m/s. Data collection utilized the OpenPose library to identify pedestrian body and facial poses, a 2D laser for distance estimation, and wheel encoders for velocity tracking. This allowed for the calculation of Time to Collision (TTC) and the classification of pedestrian behavior (walking vs. stopping) based on their interaction with the vehicle. The results, derived from 135 pedestrians, showed that 68.14% looked at the vehicle’s display, while 31.86% did not. Statistical analysis, including Pearson χ2 tests and unpaired t-tests, revealed no significant differences in pedestrian behavior across the different display conditions. Pedestrians crossed in front of the AV regardless of whether the screen showed green/open eyes or red/closed eyes. Furthermore, there were no statistically significant differences in the distance to the vehicle or the TTC at the moment of crossing between the baseline and the active display conditions. The study also found that the absence of eye contact did not significantly affect safety metrics. The authors attribute these findings to the low speed of the vehicle, which likely reduced perceived danger, and suggest that vehicle speed and distance are more decisive factors for pedestrian behavior than external visual cues. The study concludes that visual communication cues are not necessarily required for safe interaction in shared spaces governed by informal traffic rules, provided the vehicle operates at low speeds. The null hypothesis was not rejected, indicating that the specific visual signals tested did not significantly influence pedestrian yielding or crossing decisions. These findings align with prior research suggesting that vehicle motion parameters are more critical than interface displays. The authors recommend future work focus on auditory cues and additional sensors to further enhance interaction safety.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• ehmi external hmi
• vru facing ehmi
• driver vru interaction