Comparing robots with different levels of autonomy in educational setting

Abstract

Robots' ability to learn and show autonomous/intelligent behavior is expected to bring a breakthrough in usage of robots in education and assistive technologies. We compared a fully remotely operated robot (e. g. a robot with low autonomy) with one that could recognize cards and develop a playing strategy (i.e. highly autonomous) in a quartet game. We tested whether children perceive the robot in both conditions differently. Using a within-subject design, fourteen typically developed children played with a robot with high or low autonomy. The results show that both robots were evaluated equally engaging for the children. However, the introduction of more autonomy in robot’s behavior and interaction increased the time that the educator or therapist can pay attention to the child. Consequentially, the perceived usefulness of the robot and the triadic interaction between the robot, child and educator or therapist were considerably improved from the perspective of the educator.