Ghost on the Windshield: Employing a Virtual Human Character to Communicate Pedestrian Acknowledgement and Vehicle Intention

Rouchitsas, Alexandros; Alm, Håkan · 2022 · Crossref

DOI: 10.3390/info13090420

archive: archived pipeline: cataloged verified

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

Summary

This paper addresses the communication gap that will arise in autonomous vehicle (AV) traffic, where drivers will be unable to provide pedestrians with informal cues like eye contact or gestures. Pedestrians currently rely on these driver-centric cues alongside vehicle kinematics to make street-crossing decisions. To mitigate safety risks and improve traffic flow, the authors propose an external human–machine interface (eHMI) concept employing a Virtual Human Character (VHC) displayed on the vehicle’s windshield. The VHC communicates pedestrian acknowledgement via gaze direction and vehicle intention (yielding, non-yielding, or cruising) via facial expressions. The study aims to evaluate whether this anthropomorphic interface allows pedestrians to intuitively and accurately interpret vehicle intentions without prior training. The researchers conducted a monitor-based laboratory experiment using a 2 × 2 × 7 within-subject design. Participants viewed 3D animated sequences of male or female VHCs exhibiting either direct or averted gaze, combined with seven facial expressions: angry, surprised, head shake, neutral, cheek puff, smile, or nod. In each trial, participants imagined they were pedestrians intending to cross a one-way street at an uncontrolled location and indicated via mouse click whether they would cross in front of the approaching AV. The study manipulated gaze direction to simulate scenarios involving a single pedestrian (direct gaze) or two co-located pedestrians (averted gaze, implying communication with a neighbor). No explanation of the interface’s logic was provided to ensure intuitive responses. The results demonstrated that specific implementations of the VHC concept were highly effective. For communicating non-yielding intention, angry expressions, surprised expressions, and head shakes successfully deterred participants from crossing. For cruising intention, the cheek puff expression effectively signaled that the vehicle would not yield. For yielding intention, nodding successfully encouraged participants to cross. The interface proved effective in ensuring safety for both single pedestrians and pairs of co-located pedestrians without compromising traffic flow. The study found that dynamic expressions like the cheek puff were superior to static neutral expressions for indicating cruising, as they emphasized interaction potential. The significance of this work lies in its contribution to the development of intuitive, culture-transcending eHMIs for autonomous vehicles. By leveraging universal social cues such as gaze and facial expressions, the proposed interface avoids the language barriers associated with textual messages and the liability issues associated with instructional pictorial symbols. The findings suggest that anthropomorphic interfaces, when designed with high pictorial realism and appropriate positioning, can effectively support pedestrian safety and traffic efficiency. The study highlights the potential for VHCs to handle complex social interactions, such as communicating with multiple pedestrians simultaneously, thereby promoting public acceptance of AV technology.

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_with_issues.

Topics

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

• vru facing ehmi
• ehmi external hmi
• driver vru interaction
• pedestrian behavior perception
• signaling behavior