Influence of Variable Environments and Character-Specific Design on Perception of Virtual Robots with Affective Labels

Abstract

As robotic systems move out of the factory, it is essential that they adapt to changing environments. Distinct contextual settings may cause users to have a different perception of these systems. This paper explores the relationship between context and expression by looking at various combinations of robot characters and environments in a virtual setting. The robots are designed after characters from pop-culture. Physical traits such as color, eye shape, and motion profiles are abstractly analyzed and infused onto the robot. Six different robot characters and six different environments are used to see if individuals perceive each robot differently in varying environments. Additionally, the effect of character-specific priming as a context generator is also surveyed. It is found that the character with an affect label with positive valence was rated more favorably than characters with negative valence affect labels in positive valence environments and vice versa is true for environments with negative valence associated. Qualitative feedback from participants provides a meaningful description for each rating choice. This study shows the importance of contextual realizations while designing robotic systems that can be considered in future work looking at user perception.