Final Report Summary - ARBI (Adapting robot behaviour based on interaction) The main goal of the project has been the application of cognitive sciences to robotics to improve the interaction between human and robot. Basically, humans expect the robot to respond in the same way another human would do. Therefore, we believe that providing robots with cognitive capabilities is essential to enhance interaction. Moreover, another crucial component for human-robot interaction is the robot's knowledge of the world. How to represent it and how to use it to efficiently interact with the human partner? Different knowledge representation between robots and humans does not allow direct communication between them. Thus, providing the robots with mechanisms that will allow them to somehow transform their internal representation to human-like representation is crucial. The project has contributed in the following areas of human-robot interaction: (a) symbolic knowledge storage and reasoning based on an ontology; (b) geometric knowledge storage and reasoning; (c) perspective taking analysis; (d) shared attention mechanisms; (e) theory of mind modelling; (f) reference grounding during interaction; (g) verbal interaction interpretation. All these contributions have been successfully implemented and integrated in a single architecture. We have designed and tested several different scenarios where each one is oriented to show the capability of the different components. The work in this project has produced several publications, where on of it has received the CoTeSys Award, and have been presented in different international conferences and workshops. Impact The project has addressed one of the main strategic interests to European society indicated in the information and communication technology (ICT) Challenge 2 work programme of the Seventh Framework Programme (FP7): '... systems that are more aware of what is going on around them, and will be able to learn, reason and interact with people more naturally' '... In order to meet this challenge, a mix of innovative scientific theory and technology are needed, based on natural and artificial cognition, in conjunction with new systems design and engineering principles and implementations for machines, robots and other devices which are robust and versatile enough to deal with the real world and to behave in a user-friendly and intuitive way with people in everyday situations.' The work addressed in this project has proved to be essential for achieving situation assessment, i.e. gather information from its environment and reason about it to create its own beliefs about the world and the other agents (humans) in it. This is a fundamental capability that robots must be provided with in order to interact with their humans partners. Moreover, one of the publications within the project has been awarded by CoTeSys, proving the positive impact of the project in the research community. Documents connexes Final Report - ARBI (Adapting robot behaviour based on interaction)
