Diseño, construcción, y validación de una plataforma robótica asistencial, social y de servicios basada en realidad aumentada

Zusammenfassung

Advances in medicine, sanitation and hygiene during the last century, have increased human longevity, and more people are required to preserve older adults' basic expectations of dignity, pride and self-worth. The field of robotics has emerged as a promising area of research in order to help in maintaining an active lifestyle during old age, and for improving the quality of healthcare given by caregivers and family members. This dissertation presents our contribution to five topics involved in socially assistive robots: robot manufacture, robot control, context-awareness, social navigation and non-verbal interaction using augmented reality. Regarding robot manufacturing, we design, build and test a research robot able to support classic assistive topics in robotics. We show that it is possible to manufacture a medium-size robotic structure mounted on a wheeled mobile base in less than 5 hours. In order to accomplish this goal we apply a mixed manufacturing method based on rapid prototyping techniques like laser cutting, press fitting and joinery on a poplar plywood material. This embodiment allows us to integrate and test all topics previously enumerated. Relating to the second topic, this thesis proposes a hybrid control architecture that integrates inner motivational drives. It has been designed and developed to control the robot and to generate resilient robot behaviours under dynamic environments, as well as to engage the users during an activity. Attending the third topic, we describe how to develop a context-awareness component able to recognize the user activity using classical perspectives: localization, dialogue, perception, and temporal contexts. Our contribution takes advantage of environmental acoustic signals perceived at home in order to enhance the user-activity classification procedure of the context-awareness component. Our fourth and fifth topics have one element in common, the Augmented Reality (AR). We propose the use of this technology for the sake of enhancing the experience between humans and socially assistive robots. Consequently, our robot is able to offer augmented information locally and remotely. With the concept of locally in mind, we afford the social navigation problem using AR. We propose a navigation approach to offer an AR system effectively. Our approach investigates the use of social conventions to provide this kind of non-verbal feedback to users. It also introduces a solution based on proxemics principles. For the purpose of analyzing the AR effects in the interaction process, we propose the use of AR applications for assistive tasks from two points of view: in a local manner, offering a medication control system onboard our robot; and in a remote manner, providing a method to improve teleoperation in non-experienced users. We conduct a set of experiments to analyze all the proposals presented in this dissertation. The results of these experiments point out the willingness of older adults to accept new technologies in their homes. They also show that AR is a key element for improving the user experience with robots in assistance environments.