Social Human-Robot Interaction is Personalized Interaction

Previous research has indicated that robots that show social behavior when interacting with humans are better accepted. In this paper we argue that the robot behavior does not only need to be social, but also tailored to individual user characteristics, in order to be experienced as positive by the user in long-term use. To realize such personalized human-robot interactions, we need to identify differences among users, represent them in a user model and develop a number of design variants for the robot behavior. We propose to augment the conventional human-centered design process to focus on obtaining the relevant user information and creating matching design variants. We describe the different phases of the HCD4Personalization and illustrate them with examples from the NIKA project which is aimed at developing acceptable and positive interaction strategies for social robots that support older adults in their homes.