Jan Schulten, André Helgert, Alexander Arntz, Carolin Straßmann, Sabrina C. Eimler

• The application of Virtual Reality (VR) in Human-Robot Interaction (HRI) research is growing due to its capacity to create adaptable yet controlled study environments. While VR often achieves data validity comparable to real-world studies, it faces limitations when physical interaction with robots is required. Mixed Reality (MR) offers a potential solution by enabling interactions with virtual robots in real-world settings, allowing participants to engage with realistic, interactive simulations that adapt to physical environments. Given the limited use of MR in HRI studies and the need for further understanding of its effectiveness in generating transferable results, this study examines MR's potential as a tool for simulating social-robot interaction. We conducted a study involving 21 participants interacting with a virtual robot performing navigation tasks in a real library and assessed self-efficacy, presence, technology acceptance, perceived realism, and social characteristics of the robot. The results show that most participants perceived the interaction with the robot as successful, the engagement was high and the robot interaction with MR provided excellent usability. These results suggest that MR has potential as a research and training tool for simulating human-robot interaction in navigation tasks within public spaces.