Our mission is to develop robotic systems which are user-friendly and behave as cognitive partners for the humans. Based on that, we dedicate our efforts to developing cognitive functionalities for robots to endow them with versatile skills such as perception, decision making, planning, motion coordination and learning from experiences in a multi-human multi-robot framework.
Research Activities - an overview
- Human-robot collaboration
- Human-robot effort sharing
- Role allocation
- Motion coordination
- Incremental learning for haptic assistance
- Human intention estimation
- Action coordination in robot-robot collaboration
- Learning manipulation of articulated objects
- Probabilistic rigid motion estimation
- Robust and high-fidelity teleoperation
- Deformable-object manipulation
- Learning task constraints and force regulation skills in compliant manipulation.
- Two Kuka lightweight arms.
- Murola robots: 4 similar customized robots lie in our lab with the following characteristics:
- human-size height
- omni-directional mobile platform
- laser range finders at ground level
- two antropomorphic arms with 7-DoF, equipped with force sensors
- different attachable end-effectors such as two- and three-finger grippers
- emotion display head 'EDDIE' on an actuated neck and stereo camera head on pan/tilt unit respectively
- Li-Ion batteries and up to three rack-mounted computers
- Pan-tilt unit and Kinect sensor
- Additional off-board sensing systems help the robots to find their way:
- cluster of over-head cameras mounted on the ceiling frame &&
- marker-based motion tracking system
Within MuRoLa the software framework ARCADE (Architecture for Real-time Control and Autonomous Distributed Execution) is developed. ARCADE serves as a common platform for robotic research.
Human-robot physical effort sharing and role allocation;
Tight couplings between agents in physical interaction, varying contributions of the agents and potentially different action plans make intuitive physical cooperation between humans and robots a challenging research topic. Based on our system-theoretic approach on effort sharing between multiple agents, we synthesize role behaviors in terms of physical contribution a robot can attain.
Human robot dynamic movement coordination;
Coordination through synchronization occurs frequently in human behavior and seems to be a stable mediator between humans. To investigate inter-human movement coordination, we apply task paradigms which require goal-directed arm movements.
Incremental learning and control for haptic assistance;
One of the key requirements for intuitive physical human-robot cooperation is the ability of the robot to predict human behavior. Learning autonomously tasks executed together with a human partner allows the robot to anticipate human behavior reducing his work load. This active robot contribution must consider the inherent uncertainties that a prediction involves in order to become adaptive during this continuous decision making process.
Human intention estimation for safety.
Skill transfer and learning by demonstration for compliant manipulation.
We investigate transfer of manipulation skills from human experts to robots for autonomous manipulation of deformable objects. Mathematical models of deformable objects are investigated which enable robots to plan the manipulation task in advance on the one hand and refine the task execution on the other.
1. D. Althoff, O. Kourakos, M. Lawitzky, A. Mörtl, M. Rambow, F. Rohrmüller, D. Brscic, D. Wollherr, S. Hirche, M. Buss , A Flexible Architecture for Real-time Control in Multi-robot Systems , Proceedings of the Third International Workshop on Human Centered Robotic Systems '09 , Cognitive Systems Monographs , Springer, 2009.
2. F. Rohrmüller, O. Kourakos, M. Rambow, D. Brščić, D. Wollherr, S. Hirche, M. Buss, Interconnected Performance Optimization in Complex Robotic Systems, Proc. of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2010.
3. M. Rambow, F. Rohrmüller, O. Kourakos, D. Brščić, D. Wollherr, S. Hirche, M. Buss, A Framework for Information Distribution, Task Execution and Decision Making in Multi-Robot Systems, IEICE Transactions on Information and Systems, 2010, vol. E93-D: 6, p. 1352-1360.
4. M. Althoff, D. Althoff, D. Wollherr, M. Buss, Safety Verification of Autonomous Vehicles for Coordinated Evasive Maneuvers, Proc. of the IEEE Intelligent Vehicles Symposium, 2010.
5. M. Lawitzky, A. Mörtl, S. Hirche, Load Sharing in Human-Robot Cooperative Manipulation, 19th IEEE International Symposium on Robot and Human Interactive Communication, 2010.
6. D. Althoff, M. Althoff, D. Wollherr, M. Buss, Probabilistic Collision State Checker for Crowded Environments, IEEE International Conference on Robotics and Automation, 2010.