Marvin investigates the approaches of Artificial Intelligence for robotics and intelligent systems. Robots performing complex tasks in stochastic environments, such as helping humans at home or collaborating with humans in industries require cognitive skills to interpret sensor data, understand scenes, select and perform actions, deal with failures and interact with humans. We devise new methods of perception, decision and interaction for robots based on AI planning and machine learning techniques. One of the main focuses of the group is the integration of these techniques into different robotic platforms and architectures.

Our group addresses 3 scientific challenges:

• Challenge 1: Develop explainable, provable and human acceptable decision mechanisms for autonomous robots
• Challenge 2: Develop intelligent perception to model the environment and to be integrated in the robot decision-making process
• Challenge 3: Develop social robots that can be considered as intelligible, credible and acceptable for humans