Design challenges for long-term interaction with a robot in a science classroom

This paper aims to present the main challenges that emerged during the process of the research design of a longitudinal study on child-robot interaction for science education and to discuss relevant suggestions in the context. The theoretical rationale is based on aspects of the theory of social constructivism and we use the collaborative inquiry as a framework to examine children's learning process who interact with a robotic learning companion. We identify two main challenges; (i) the development of robust on-demand systems for long-term interaction; and (ii) the design of developmentally appropriate scaffolding in embodied, semi-structured learning tasks. To address these challenges, we suggest (i) the development of a system for the detection of child's intention for interaction in the context of a classroom and (ii) the design of sensorized learning materials for the support of developmentally appropriate embodied learning experience.