The RoboEarth project targets at building a "World Wide Web for Robots" and covers different aspects like the generation and execution of task descriptions, sensing, learning, a central web knowledge base; TUM is responsible for developing methods for representing and reasoning about the exchanged knowledge.

The central RoboEarth knowledge base contains descriptions of actions (called "action recipes"), objects, and environments. These pieces of information have been created by different robots with different sensing, acting and processing capabilities. Therefore, all of them have different requirements on capabilities a robot must have in order to use them. The developed language thus provides methods for matching these required capabilities against those available on the robot. Each robot has a self-model consisting of a description of its kinematic structure, including the positions of sensors and actuators, a semantic model that describes the meaning of the robot's parts (e.g. that certain joints form a gripper), and a set of software components like object recognition systems. We developed the Semantic Robot Description Language to describe these components and the capabilities they provide, and to match them to the requirements specified for action recipes.

More information can be found on the RoboEarth project homepage.

This project has received funding from the European Union Seventh Framework Programme FP7/2007-2013 under grant agreement number 248942 RoboEarth.