Robot's individual emotion generation model and action coloring according to the robot's personality

摘要

In order to achieve a believable interaction between humans and robots, synthesizing the reasonable and natural emotion is important. Furthermore, improving the adjustability of parameters' values in the model is essential for the designer of a robot. In this paper we developed a hybrid emotion generation architecture and defined the generalized context inputs of emotion generation model for convenient use by a robot designer. We also developed the personality model based on the Big-Five personality factors to generate the various emotions and action colorings. The personality model affected not only the intensity of the generated emotion, mood state, and emotion decay, but also the action selection of the robot. So, for each circumstance, the robot can show the distinct emotional state and modified actions, which we call action colorings. The developed emotion generation system is integrated into a robot which responds to the user's calling, and manages the schedule of the human.