Turn-Taking Intention Recognition using Multimodal Cues in Social Human-Robot Interaction

Turn-taking is an essential social skill for human communication. The robot needs to recognize the end of turn for timely response to the user with little delay in human-robot interaction. In this paper, we propose a turn-taking intention recognition system that determine the timing of turn-taking using multimodal cues in social Human-Robot Interaction (sHRI). In order to evaluate the turn-taking intention recognition system, we collect multimodal data set and conducted experiments. To that end, we designed a human-robot interaction scenario including turn-taking and conducted an experiment with 30 participants using the humanoid robot NAO. In experiments, we validate recognition models trained multimodal dataset by machine learning methods.