Effect of a Learning Support Model that Provides Autonomous Learning Support in a Teacher-Type Robot Based on the Learner’s Perplexion State

In recent years, the introduction of ICT education has become active, and research on educational support robots has been attracting attention, especially in this field. However, it has been reported that the conventional educational support robots, which provide learning support through button operations by learners, cause excessive support demands from learners. To solve this problem, in this study, a perplexion estimation method was proposed that estimates the perplexed state of learners from their facial expressions through deep learning. Furthermore, an apprenticeship promotion model was constructed by combining the behavior model for providing learning support based on the cognitive apprenticeship theory and the perplexion estimation method to solve this problem. This paper investigates the effects of an educational support robot equipped with the apprenticeship promotion model for university students. The results of the subject experiment confirmed that the robot using this model provides the same learning effect as the conventional robot that provides learning support by button press. In other words, this model suggests that it is possible to accurately estimate the perplexed state of learners and achieve optimal learning support timing.