A matter of perspective? The interplay of camera perspective and augmentation for remote assistance of highly automated vehicles

Kurzfassung

Remote human operators are essential for guiding Level 4 autonomous vehicles through challenging scenarios like adverse weather, but standard first-person camera views often cause object occlusion and perceptual distortions that can hinder their performance. A lab study with 37 participants examined how three different camera perspectives and sensor-data video augmentation influence operator decision-making at busy intersections under both clear and foggy conditions. The results showed that while no single camera angle was universally optimal, video augmentation effectively compensated for poor visibility in adverse weather. Ultimately, the findings suggest that human-machine interfaces for remote vehicle assistance should be context-sensitive, dynamically adapting camera perspectives and data overlays to match specific traffic situations and environmental conditions.