Autonomous Task Planning and Execution
Solving arbitrary manipulation tasks is a key feature for humanoid service robots. However, especially when tasks involve handling complex mechanisms or using tools, a generic action description is hard to define. Different objects require different handling methods. Therefore, a modular system architecture has been developed to autonomously solve manipulation tasks from the object point of view.
Functional Object Classes:
- Objects categorized by functionality in a hierarchical structure
- Description of generic process models on the symbolic and geometric level via action templates
Hybrid Reasoning based on Object Knowledge:
- An object storage provides prior object knowledge
- The current world state is used as initial state for the symbolic planner
- Action templates ground the symbolic planners outcome
- Individual robot components are addressed on the geometric level
Selected Publications
Daniel Leidner, Christoph Borst, and Gerd Hirzinger, "Things Are Made for What They Are: Solving Manipulation Tasks by Using Functional Object Classes", in Proc. of the IEEE-RAS International Conference on Humanoid Robots, Osaka, Japan, November 2012, pp. 429–435.
Contact
Daniel Leidner
Institute of Robotics and Mechatronics
Cognitive Robotics
Münchener Straße 20, 82234 Oberpfaffenhofen-Weßling