Real-Time Vision-Based Learning for Human-Robot Interaction in Social Humanoid Robotics

This report proposes a research topic in the field of human-robot interaction. The research will focus on real-time learning paradigm and its application in social robotics. Gestures, as part of more complex actions, will provide information necessary for a social robot to infer emotional state of a human demonstrator. The research may lead to further investigation of acceptance of humanoid robots in social settings. The report gives an overview over existing literature, presenting the evolution of the real-time learning and focusing on a particular subset that deals with vision-based data acquisition in social robotics tailored to research in human-robot interaction. In addition, the report enumerates the details of the project, such as aims, objectives and analyses the requirements, proposing the research plan that will lead to development of an integrated self–organising system.