Personalized Emotional Expressions to Improve Natural Human-Humanoid Interaction

Abstract

We need to prepare robots for the shift from laboratories and industrial environments to join human residential areas. This is one of the reasons why current trends in the field of human-robot interaction are expanding into the social experience of users often involving artificial emotions. Emotional technology in its two forms – as an expression of artificial emotions of the systems and as systems capable of recognizing human emotions – contributes to the creation of personalized systems. We try to move from the knowledge about human emotional processes to implement a model of artificial emotions based on Plutchik’s theories. The emotional model is a part of the autonomous mode for the humanoid robot. The proposed system is able to adapt to the user’s expectations and thus, to evolve during the interaction process. The inputs to the system are user’s expressions of emotions detected from his/her body movements and gestures. Fuzzy logic approach is used for blending of basic emotions to primary mixed emotions. We used humanoid robot Nao as an experimental setup and the user’s expressions are mapped to the humanoid’s body. The expressions are used to communicate robot’s internal states much like non–verbal signals indicates feelings in human society. As the experiments show, the personalized expressions help users to understand the system and they improve the human-robot interaction.