The sense of smell in social robots

This paper presents an approach to the integration of the sense of smell in social robots. The Intelligence Control Group of CAR (CSIC-UPM) has been working on the design of URBANO, a tour guide robot. The URBANO´s software architecture is based in agents using SOAP proposal. Some agents offer cognitive services, such as the automatic generation of presentations, learning of concepts using internet and the emotional behavior. One of them is an ontology that contains the knowledge used by the robot. A new agent that implements the learning of odors and produces new stimuli has been developed. This stimuli generate changes in the emotions, therefore, it also produces changes in the behavior of the robot. The obtained results show how to add odors to the emotional model of the robot help it to increase its performance as a social robot.