Conceptual Motorics - Generation and Evaluation of Communicative Robot Gesture

摘要

How is communicative gesture behavior in robots perceived by humans? Although gesture is a crucial feature of social interaction, this research question is still largely unexplored in the field of social robotics. The present work thus sets out to investigate how robot gesture can be used to design and realize more natural and human-like communication capabilities for social robots. The adopted approach is twofold. \nFirstly, the technical challenges encountered when implementing a speech-gesture generation model on a robotic platform are tackled. The realized framework enables a humanoid robot to flexibly produce synthetic speech and co-verbal hand and arm gestures at run-time; in contrast to many existing models, these gestures are not limited to a predefined repertoire of motor actions. Fine synchronization of the two modalities is achieved by means of a sophisticated multimodal scheduler specifically implemented for humanoid robot gesture and speech.\nSecondly, the achieved flexibility in robot gesture is exploited in controlled experiments. To gain a deeper understanding of how communicative robot gesture might impact and shape human perception and evaluation of human-robot interaction, two experimental studies were conducted. The findings reveal that participants evaluate the robot more positively when non-verbal behaviors such as hand and arm gestures are displayed along with speech. \nSurprisingly, this effect was particularly pronounced when the robot's gesturing behavior was partly incongruent with speech. These findings contribute new insights into human perception and understanding of non-verbal behaviors in artificial embodied agents. Ultimately, they support the presented approach of endowing social robots with communicative gesture.