Real-World Effectiveness of LED eHMIs for Supporting Pedestrian Interactions with Multiple Highly Automated Vehicles

Kurzfassung

Highly automated vehicles (HAVs) promise safer, more efficient mobility, but also introduce new communication challenges for pedestrians. Traditional cues such as eye contact or gestures are absent when no human driver is present, limiting pedestrians to vehicle kinematics as their primary source of information. Yet, at low speeds, this implicit communication may be ambiguous, creating a risk of misunderstandings. This might be even worse when a pedestrian has to cope with multiple HAVs as interaction partners. Previous simulator studies have shown that LED light-based external human-machine interfaces (eHMIs) can support pedestrians when interacting with driverless vehicles by offering explicit communication signals. In our current real-world test-track study using two automated test vehicles 50 participants adopted a pedestrian's perspective in a road crossing scenario. Participants decided when to cross while two HAVs approached from opposite directions. Two eHMI conditions (no eHMI vs. LED light band) and four vehicle kinematic conditions (left HAV, right HAV, both, or neither HAV yielding) were tested. As outcome measures, crossing initiation time, perceived safety, and perceived eHMI support were collected. Results revealed that eHMI significantly reduced crossing initiation time, enhanced perceived safety, and supported pedestrians' decision-making. Via explicit communication signals, the LED band effectively complemented vehicle kinematics, mitigating confusion and boosting pedestrian's confidence. Furthermore, this study confirms eHMIs' feasibility in real-world settings, reinforcing their value in supporting pedestrian-HAV interaction. These findings underscore the importance of combining explicit eHMI signals with inherent vehicle kinematics to promote safer, more intuitive, and inclusive interactions in future automated traffic environments.