Modeling and Evaluating Emotion, Contextual Head Movement and Voices for a Social Robot Storyteller

Storytelling bears great potential as prospective application for social robots. Through their embodiment and potential for multi-modal behavior, they can enhance current storytelling media, for example books or audio books. We combined theoretical knowledge about designing storytelling robots with an analysis of human storytellers to implement the behavior of a robot storyteller. We tested the following modalities in three perception studies: emotional behavior, contextual head movements, and voice acting through different synthetic voices. Our results suggest, that an emotional robot storyteller is able to transport participants equally well as the traditional audio book, contextual head movement has to be carefully modeled to be recognized, and voice acting can improve narrative presence and can make the robot more interesting. With the present contribution, we provide guidance for modeling and evaluating the behavior of a social robot storyteller, and suggest that the careful integration of emotional display and different voices can enhance the perception of a robot storytelling experience.