The Cognitive Robot Abstract Machine
Lorenz Mösenlechner · mediaTUM – the media and publications repository of the Technical University Munich (Technical University Munich) · 2016
New mobile robotic platforms such as Willow Garage's PR2 that can operate in a human household are able to perform complex tasks such as cooking food or cleaning up. However, to solve these tasks reliably and robustly requires the integration of action execution and reasoning-based decision making. This thesis describes CRAM, a complete framework for implementing cognitive high-level robot control programs to enable robotic agents to execute their tasks more robustly, reliably and flexibly. Besides a powerful programming language, CRAM integrates different reasoning components including reasoning about what the robot did, when it did it and why, and accurate geometric reasoning to infer, for instance, from which locations an object is visible.