Empathetic robot evaluation through emotion estimation analysis and facial expression synchronization from biological information

Empathy is an important factor in human communication. For a robot to apply a matching emotion in human–robot communication, the robot needs to be able to understand human feelings. Therefore, in this study, we aimed to improve the human impression of the robot using a robot that expresses human-like expressions by synchronizing with human biological information and changing the expressions in real time. We first measured and estimated human emotions using an emotion estimation method based on biological information (brain waves and heartbeats). The three-emotion estimation methods were proposed and evaluated in the preliminary experiment. Among the three-emotion estimation methods proposed, the one that yields the highest impression rating was chosen to be used in the second experiment which was based on the emotional value in each cycle method. Then, we developed a robot that shows expressions in two patterns: (1) synchronized emotion (same emotion as subject conveyed) and (2) inversed emotion with the human. The subjects evaluated the robot’s expression from both patterns using semantic differential (SD) method while having their biological information measured based on the selected emotion estimation method from previous preliminary experiment. The evaluation by SD method and biological information results showed that when the human experienced the happiness emotion, and the robot synchronized and expressed the same emotion, this could increase the intimacy between human and robot. Here, it can be said that the impression created by the robot’s expression can be improved using biological information.