Encountering Robots in the Field: Proof of Concept and Findings from a Real-Life Eye-Tracking Study

Abstract The successful integration of robots into public spaces depends not only on their technological capabilities but also on the robots’ ability to align with human expectations (e.g., safety concerns). Therefore, understanding how people perceive robots and react to them is essential in human-robot interaction (HRI), and the initial encounter between humans and robots is especially important. In this study, we investigated people’s gaze behavior when approaching an autonomous robot initially under naturalistic field settings. Altogether, $$N = 29$$ people were invited to a public area where they encountered an autonomous cleaning robot. The criticality of the robot behavior was manipulated between participants (a robot crossing the pedestrian path vs. a robot moving parallel to the pedestrian path). Participants were asked to walk a specific pathway with eye-tracking glasses without knowing the study context to be able to assess natural, unbiased gaze behavior and behavioral reactions. The results showed that potential conflicts between pedestrians and robots lead to higher numbers of fixation and more changes in behavior. However, they did not affect crossing times. Intermediate proximity (1–3 meters) to the robot was associated with the most visual attention compared to close and far proximity. Importantly, critical scenarios did not impact the perceived safety, trust, and acceptance of robots, indicating that behavioral changes were overall successful and reduced the perceived criticality of the encounter.