Shopping with a robotic companion

In this paper, we present a robotic shopping assistant, designed with a cognitive architecture, groundedin machine learning systems, in order to study how the human-robot interaction (HRI) is changing theshopping behavior in smart technological stores. In the software environment of the NAO robot, connectedto the Internet with cloud services, we designed a social-like interaction where the robot carriesout actions with the customer. In particular, we focused our design on two main skills the robot has tolearn: the first is the ability to acquire social input communicated by relevant clues that humans provideabout their emotional state (emotions, emotional speech), or collected in the Social Media (such as,information on the customer's tastes, cultural background, etc.). The second is the skill to express in turnits own emotional state, so that it can affect the customer buying decision, refining in the user the senseof interacting with a human-like companion. By combining social robotics and machine learning systemsthe potential of robotics to assist people in real life situations will increase, providing a gentle customers'acceptance of advanced technologies.