A new kind of robots whose characteristics, objectives and operational modes drastically differ from more conventional industrial robots is gaining increased interest. This new type of robots aims to achieve a high level of flexibility, adaptability, and efficiency for acting in environments designed for humans. To meet these demands, the mobile robots need lots of information about their surroundings at different levels of abstraction, generally in real-time. In this paper a sensor concept is described containing symbolic and sub-symbolic information delivering the required data for all tasks. Based on this information the planning, navigation, and collision avoidance of the ARIADNE robots is done.

Michael Pauly, Hartmut Surmann, Marion Finke and Nang Liang. Real-Time Object Detection for Autonomous Robots. Informatik Aktuell. Autonome Mobile Systeme, 14. Fachgespräch, Springer-Verlag, S. 57-64, 1998.