The aim of this thesis is to present a novel decision-making system based on bio-inspired concepts to decide the actions to make during the interaction between humans and robots. We use concepts from nature to make the robot may behave analogously to a living being for a better acceptance by people. The system is applied to autonomous Socially Interactive Robots that works in environments with users. These objectives are motivated by the need of having robots collaborating, entertaining or helping in educational tasks for real situations with children or elder people where the robot has to behave socially. Moreover, the decision-making system can be integrated into this kind of robots in order to learn how to act depending on the user profile the robot is interacting with. The decision-making system proposed in this thesis is a solution to all these issues in addition to a complement for interactive learning in HRI. We also show real applications of the system proposed applying it in an educational scenario, a situation where the robot can learn and interact with different kinds of people. The last goal of this thesis is to develop a robotic architecture that is able to learn how to behave in different contexts where humans and robots coexist. For that purpose, we design a modular and portable robotic architecture that is included in several robots. Including well-known software engineering techniques together with innovative agile software development procedures that produces an easily extensible architecture.
© 2001-2024 Fundación Dialnet · Todos los derechos reservados