Human-robot interaction methodology: Robot teaching activity

Abstract

Research on the use of social robots in education is constantly increasing in the growing field of human-robot interaction (HRI). Consequently, it is essential to determine an appropriate methodology to test how these robots can optimally interact with students. This study specifically looks at how we can use existing knowledge from psychology, neuroscience and educational research and apply them with validity and credibility in HRI studies. We are interested in incorporating research methodologies to evaluate the performance of social robots acting as university professors in a real classroom environment. Moreover, we aim to measure three effects, a) students’ knowledge acquisition (quiz after lecture and final exam grades), b) level of enjoyment (Likert scale questionnaire), and c) level of surprize (analysis of facial expressions filmed by cameras). To identify the relationship between students’ knowledge acquisition, enjoyment, and level of surprize, we designed a series of three experiments, testing three variables: 1. one human-tutor lesson, 2. one robot-tutor lesson, 3. two robot-tutor lessons. In this paper we thoroughly explain the methods used to measuring and testing these variables based on modern and reliable sources.