Relationship between social robot proactive behavior and the human perception of anthropomorphic attributes

Various attempts have been made to research the proactive interaction behavior of robots and to apply different service scenarios from home to public areas but have not investigated social robot behavior and user perception from the anthropomorphizing design's preceptive to construct robots' interaction strategy. Our aim is to explore the relationship between proactivity in social robot behavior and a user's perception of anthropomorphic attributes. We conducted a Wizard-of-Oz study by designing five levels of proactive behaviors in a social robot. The subjects experienced each of them in the experiment and rate their perception on a scale, which includes anthropomorphic attributes such as Warmth and Mind Attribution. The subjects also provided an explicit description of their social experience regarding proactive behaviors. The result indicated that all anthropomorphic attributes of different proactivity levels have a significant difference. And the user perception of the anthropomorphic attributes in terms of Warmth can be increased with the proactive behavior levels. However, when robots do not use proactive behavior, users may perceive more Mind Attribution. Understanding a human's perception of anthropomorphic attributes of social robots with different proactive levels of robot behaviors can support HRI design in domestic robot applications.