Robot's behavior expressions according to the sentence types and emotions with modification by personality

Abstract

Expression has become one of important parts in human-robot interaction as an intuitive communication channel between humans and robots. However it is very difficult to construct robot's behaviors one by one. Developers consider how to make various motions of the robot easily. Therefore we propose an useful behavior expression method according to the sentence types and emotions. In this paper, robots express behaviors using motion sets of multi-modalities described as a combination of sentence types and emotions. In order to gather the data of multi-modal motion sets, we used video analysis of the actress for human modalities and did user-test for non-human modalities. We developed a behavior edit-toolkit to make and modify robot's behaviors easily. And also we proposed stereotyped actions according to the robot's personality for diversifying behavior expressions. Defined 25 behaviors based on the sentence types and emotions are applied to Silbot, a test-bed robot in CIR of Korea, and used for the English education.