 |
| Detailed information |
| Original study plan |
Master's programme Mechatronics 2026W |
| Learning Outcomes |
Competences |
| Students develop a research-oriented understanding of modern robot learning methods, which describe learning approaches to generate goal-directed behaviors on robotic systems. The class focuses on reinforcement learning, imitation learning, simulation-to-reality transfer, foundation models for robotics (specifically diffusion and transformer-based models), and bio-inspired as well as morphology-aware robotics. They can formulate robotic control problems within contemporary learning frameworks, analyze learning-based control architectures
for dynamic and manipulation tasks, and understand how large-scale generative and transformer models can be integrated into robotic systems. Students can critically evaluate state-of-the-art research in robot learning and
independently design learning-based solutions for high-dimensional, dynamic, and real-world robotic platforms.
|
|
| Skills |
Knowledge |
|
|
|
| Criteria for evaluation |
|
| Language |
English and French |
| Changing subject? |
No |
|
