External Human-Machine Interfaces on Automated Vehicles: Effects on Pedestrian Crossing Decisions

Area of Science:

• Human-Computer Interaction
• Automotive Engineering
• Transportation Safety

Background:

• Pedestrian safety and efficiency in automated vehicle interactions are critical.
• External human-machine interfaces (eHMIs) on automated vehicles may improve these interactions.
• Clear communication of vehicle intent is essential for safe road sharing.

Purpose of the Study:

• To investigate the impact of various external human-machine interfaces (eHMIs) on pedestrian crossing intentions.
• To evaluate how different eHMI types and timings affect pedestrians' perceived safety.
• To determine the most effective eHMI for enhancing pedestrian-automated vehicle interactions.

Main Methods:

• A virtual reality simulation using head-mounted displays exposed 28 participants to urban road scenarios.
• Vehicle behavior (yielding/non-yielding), size, eHMI type (none, brake lights, Knightrider, smiley, text), and timing were manipulated.
• Participants continuously reported their 'feel-safe' status via a button, calculating percentages for each condition.

Main Results:

• Pedestrian 'feel-safe' percentages increased with eHMIs (brake lights, Knightrider, smiley, text) for yielding vehicles compared to baseline.
• For non-yielding vehicles, eHMI presence/type did not significantly alter 'feel-safe' percentages, though larger vehicles reduced it.
• Text-based eHMIs demonstrated immediate understanding without a learning curve, unlike other tested visual cues.

Conclusions:

• External human-machine interfaces (eHMIs) enhance the efficiency of pedestrian-automated vehicle interactions.
• Textual eHMI displays are perceived as the least ambiguous, improving clarity for pedestrians.
• This research supports the integration of communication systems in automated vehicles for safer road usage.