Skip to main content Skip to main navigation


Non-verbal Sound Detection by Humanoid Educational Robots in Game-Based Learning. Multiple-Talker Tracking in a Buzzer Quiz Game with the Pepper Robot

Ilona Buchem; Rezaul Tutul; André Jakob; Niels Pinkwart
In: Richard Balogh; David Obdr¸álek; Eftychios Christoforou. International Conference on Robotics in Education (RiE). Pages 185-196, LNNS, Vol. 747, ISBN 978-3-031-38454-7, Springer Nature Switzerland, 4/2023.


While most research and development in social human–robot interaction, including interactions of learners with educational robots, has focused on verbal communication through speech (synthesis and recognition), research on non-verbal signals has just started to gain attention. Non-verbal sounds such as clapping, laughter, and mechanical sounds are often overlooked elements in design and research in social human–robot interaction. This paper presents the design of an educational buzzer quiz game for groups of learners, which is facilitated by the humanoid educational robot Pepper, detecting non-verbal sounds as part of the game-play. The paper describes the scenario and a prototype of a robot-supported buzzer quiz game in classroom settings, outlines the challenges in the real-time detection of a mixture of sounds, and presents the approach to real-time auditory and visual multiple-talker tracking, which has been applied in the buzzer quiz game with the Pepper robot.