Learner Modelling and Automatic Engagement Recognition with Robotic Tutors

In this paper we discuss the initial steps of the design of a computational model capable of automatically detecting user engagement with an artificial robotic tutor. Generally, our model will be used with a robotic tutor to support and enhance the experience of the learner by regulating pedagogical and empathetic interventions in a timely manner. We describe the requirements of the learner model in order to understand the state of the learner and facilitate the learning progress. Additionally, we propose the initial steps of the design of a suitable scenario for the learning task activity to allow the model to be tested on actual class material from UK curriculum based on teachers' feedback. Finally, we discuss an initial plan to build the model with real user data from a pilot study based on Wizard-of-Oz (WoZ) conducted in a real classroom environment.