Human trust and emotion in the context of collision avoidance with an autonomous mobile robot: An investigation of predictability and smoothness in virtual reality

Abstract

The integration of mobile robots into human environments requires that they behave in trustworthy ways. Trust develops through interaction, and erratic movements can easily break it. Previous research examined motion predictability and smoothness separately, but their combined effect remains unclear. We investigated how predictability (consistent versus random) and smoothness (gradual versus abrupt) influence human trust and emotion during collision avoidance in virtual reality. Twenty-six participants encountered a robot across repeated trials. We measured subjective valence, arousal, and trust, along with skin conductance responses. Predictability dominated the results. Consistent robot behavior led to increased trust and positive feelings over time. Conversely, unpredictable behavior kept trust low. Smoothness acted mainly as a moderator for arousal; smooth paths reduced the stress caused by unpredictable moves. We also found that prolonged proximity in gradual movements raised physiological arousal, even if users reported feeling calm. These findings suggest predictability drives social acceptance more than smoothness. In order to promote trust, robot designers should prioritize consistent, learnable behaviors. We examined AMR motion predictability and smoothness in VR collision avoidance. Predictable motion significantly improved human trust and valence over time. Unpredictable and abrupt behavior increased physiological arousal. Smoothness modulated arousal primarily when behavior was unpredictable. Predictability is more critical than smoothness for social acceptance of AMRs.