It's NAO or Never! Facilitate Bonding Between a Child and a Social Robot: exploring the possibility of a robot adaptive to personality
No Thumbnail Available
In this thesis we investigate whether it is effective to let a social robot adapt itself to personality (extroversion) to aid bonding between a robot and a child. After a short exploratory study, we designed an experiment where 14 children aged 10 played a mimicking game for 10 minutes with two different robots. One robot moved slightly slower and gave nurturing feedback (the `introvert' robot) and the other moved slightly faster and gave challenging feedback (the `extrovert' robot). Our first question was whether the children noted the differences, and the second was if the children had a preference related to their own personality. To answer these questions we assessed personality (extroversion) of the children with a subset of the Big Five Questionnaire for Children (BFQ-C), furthermore we adopted both subjective and objective metrics to analyze the interaction regarding fun, bonding and trust. Further issues that need to be addressed are whether the robot is able to infer extroversion from the behavior of the child and secondly whether the robot is able to learn meaningful adaptation rules, some preliminary work is done to provide guidelines to address these issues. The main conclusion is that the children very much liked playing the game with both of the robots. There was no difference found between the two robots, it is possible that they are not distinguishable enough, and the experiment lasted not long enough to pass the novelty effect and reveal the subtle differences. There was no preference related to the child's personality. But the score on extroversion was densely distributed which raises some questions on personality assessment with children of this age in general. Our conclusion is that our robot already elicited a lot of fun, regardless of personality adaptation. To further facilitate bonding there are a lot of other aspects which need attention first, for example the gender of robot or child or the development of optimal scenarios and dialogues. We have to make the giant leaps in child-robot interaction elsewhere before returning to personality.
Faculteit der Sociale Wetenschappen