Adaptable semi-autonomy in personal robots

Abstract

Personal robotics is widely recognized as a major challenge for current robotics research. Robots assisting humans and closely interacting with them have to meet acceptability requirements which in turn leads one to reconsider the concept of autonomy in robotics. The paper presents an abstract analysis of possible levels of semi-autonomy in personal robots and illustrates a case study in which adaptable semi-autonomy is implemented and experimentally validated. Explanation modules for human-robot communication in the planning phase, and user modeling techniques, that allow the system to adapt to its user's needs and preferences, are proposed as ways to achieve adaptive semi-autonomy.