Método centrado no usuário para construção de robôs autônomos visando a interação com humanos

Abstract

The robotics scene in the world has changed over the last few years. Before the robot environment had its focus only on the industry. Now there is a greater emphasis on so-called service robots. The International Federation of Robotics (IFR) defines service robots as any robot out of an industrial setting. They are in areas such as agriculture, hospitals, hotels, offices and residences. In this way, the need for interaction between robots and humans becomes inevitable. Thus, researchers have dedicated to creating projects concerned with social humanrobot interaction. The human-robot interaction works to guarantee more quality in interaction. It uses techniques of user experience, control and artificial intelligence. Human-robot interaction discussed in the literature focus on small parts of the project. Some examples are: touch control, naturalness of gestures and movements, deviations of obstacles, recognition of people, among others. These parties are more concerned with the behavior and skills of the robot than with the user. There is a lack of a systemic method that enables the documentation, evolution and maintenance of autonomous service robots. Besides to the systemic method, one that is user centered is important given the joint activity to the human being. This thesis presents a user-centered method for constructing an autonomous robot. It focus on human-robot interaction. The proposed method is applied in a case study based on a domestic scenario, where the robot interacts through voice and has autonomous navigation. The Personas technique is used to identify the user profile. A Bayesian classifier is presented as a way to illustrate the proposal of a part of the decision making stage of the robot. The results show the importance of method, by the design evolution. Also, maintenance that were necessary during the tests in the application of the case study.