Basic Study of Estimating Learners’ Confusion for Educational Support Robot

Abstract

Recently, educational support robots attract attention. In this study, we focus on teacher-type robots which behave like teachers. In previous studies, teacher-type robots were effective in collaborative learning with learners. However, robots provide learning support when a question is presented or when learning support is requested by a learner's button operation. The previous studies did not reported that robots autonomously provide learning support. Therefore, we develop a teacher-type robot which can autonomously provide learning support by estimating learners' confusion. This paper proposes a method for estimating learners' confusion from learners' facial expression in a framework of the ensemble learning.