V okviru laboratorija CoBoTaT razpisujemo diplomske in magistrske teme na področjih robotike in strojnega učenja. Teme so lahko odskočna deska za nadaljnje sodelovanje. CoBoTaT je novo ustanovljen laboratorij, katerega cilj so raziskave na področju učinkovitega sodelovanja med človekom in robotom. V ta namen je laboratorij opremljen s sodobnimi merilnimi sistemi za zajem človeškega gibanja in roboti, ki so lahko v kontaktu s človekom. Razpisane magistrske naloge se lahko opravijo na robotih Panda Franka Emika, Kuka LWR in humanoidnem robotu TALOS.
Študentje lahko temo diplomskega in magistrskega dela predlagajo tudi sami.
1. Human motion prediction for hand-overs
In human-robot physical collaboration, robots currently face a shortcoming due to their limitations in observing human motion. This further results in inefficient collaboration and unergonomic interaction. To overcome this shortcomings one of the key enabling technologies for this, is the ability to hand over objects between partners (human and robot) using only non-verbal communication, such as gestures and body pose. In this project, we plan to use Compliant Movement Primitives (CMP) to represent the handover motions, and design an algorithm that can predict future human movements. To goal is to detect human hand pose with good enough precision in time and space to enable human like hand-overs of robots. Human will be detected using a Microsoft kinetic system. Evaluation on a real or simulated robot will be used to validate the prediction and task performance.
In robotics, the manipulability ellipsoid is the geometric interpretation of the scaled eigenvectors resulting from the singular value decomposition of the jacobian that describes a robot’s motion.
Requirements:
- Good programming skills in Matlab
- Basic understanding of robot kinematics and control
- Knowledge of Linux / Robot Operating System (ROS) / C++ is a plus
Literature:
- Deniša, M., Gams, A., Ude, A., & Petrič, T. (2016). Learning compliant movement primitives through demonstration and statistical generalization. IEEE/ASME transactions on mechatronics, 21(5), 2581-2594.
2. Ergonomic Control of Human-Robot Handover
In human-robot physical collaboration, robots currently face a shortcoming due to their limitations in observing human motion. This further results in inefficient collaboration and unergonomic interaction. To overcome this shortcomings one of the key enabling technologies for this, is the ability to hand over objects between partners (human and robot) using only non-verbal communication, such as gestures and body pose. In this project, we plan to use human manipulability measures to optimize the robot position of the co-manipulation task in the work-space. In such configuration we will reduce human effort by minimizing the effects of an external load in human body joints. Human will be detected using a Microsoft kinetic system. Evaluation on a real or simulated robot will be used to validate the prediction and task performance.
In robotics, the manipulability ellipsoid is the geometric interpretation of the scaled eigenvectors resulting from the singular value decomposition of the jacobian that describes a robot’s motion.
Requirements:
- Good programming skills in Matlab
- Basic understanding of robot kinematics and control
- Knowledge of Linux / Robot Operating System (ROS) / C++ is a plus
Literature:
- L. Peternel, W. Kim, J. Babič and A. Ajoudani, “Towards ergonomic control of human-robot co-manipulation and handover,” 2017 IEEE-RAS 17th International Conference on Humanoid Robotics (Humanoids), Birmingham, 2017, pp. 55-60.
- Goljat, R., Babič, J., Petrič, T., Peternel, L., & Morimoto, J. (2017). Power-Augmentation Control Approach for Arm Exoskeleton Based on Human Muscular Manipulability. In 2017 IEEE International Conference on Robotics and Automation. IEEE.
3. Using quadratic programming for redundancy resolution
In this project the focus will be given on solving redundancy resolution using online optimization technics, like quadratic programing. The aim is to optimize the pose of redundant robots with compliant joints to achieve desirable end-effector stiffness or manipulability. Evaluation on a real or simulated robot will be used to validate the prediction and task performance.
A manipulator is termed kinematically redundant when it possesses more degrees of freedom than it is needed to execute a given task. Redundancy can be conveniently exploited to achieve more dexterous robot motions. The inverse kinematic problem is of particular interest in this case since it admits infinite solutions.
Requirements:
- Good programming skills in Matlab
- Basic understanding of robot kinematics and control
- Knowledge of Linux / Robot Operating System (ROS) / C++ is a plus
Literature:
- Albu-Schaffer, A., Fischer, M., Schreiber, G., Schoeppe, F., & Hirzinger, G. (2004, September). Soft robotics: what cartesian stiffness can obtain with passively compliant, uncoupled joints?. In Intelligent Robots and Systems, 2004.(IROS 2004). Proceedings. 2004 IEEE/RSJ International Conference on (Vol. 4, pp. 3295-3301). IEEE.