Eliciting ideal tutor trait perception in robots Pinpointing effective robot design space elements for smooth tutor interactions

Abstract

To approach the physical design of a tutor robot, we obtained 3 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">rd</sup> to 5 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">th</sup> grade children's evaluations of the relative importance of tutor traits, and of which robot design categories they perceive as most embodying top tutor traits, in an initial exploratory interview study. Results indicate that prototypical “Mechanoid” (humanoid or robotic) designs and animal-shaped designs (whether animal-like in outer cover, or more visibly a robotic animal) are superior to object-based designs (desktop objects or geometric shapes). Furthermore, children's perception of top tutor traits can be better pushed by an animal-shaped design than by a Mechanoid design, with less risk of unintended Uncanny Valley effects. Implications for designing tutor-robot embodiments toward buttressing children's expectations of an ideal tutor and for facilitating interactions, and future research directions are discussed.