How wild is wild? A taxonomy to characterize the ‘wildness’ of child-robot interaction

When thinking about Child-Robot Interaction (CRI) in the ‘wild’ or natural settings, many ideas come to mind, such as a home or a school that involve chaotic settings with autonomous robotic devices and people that are freely interacting with them. However, there certainly are degrees of ‘wild’, and different experimental settings can have varying levels of control in place. It would be helpful to have a common framework to interpret and identify the many different influencing factors or levels of control surrounding CRI experimentation. Having a framework to help towards standardizing evaluation of CRI studies would benefit researchers wishing to identify or plan the varying dimensions present in CRI experimentation. This paper presents a simple taxonomy to characterize the ‘wildness’ factors in CRI over two main dimensions (Participant and Robotic) that can effect the overall outcome of such studies. The use of this taxonomy is illustrated by its application to current CRI research. Specifically, we use it in reflection to rate six of our CRI trials that have been conducted over a ten year period. From the classification of these studies, a general view of our work so far is outlined and new research perspectives are identified. The application of the taxonomy is also validated by reviewing a selection of other CRI studies.