M.Sc. Thesis Offer: Dynamic Legged Locomotion using Multi-Modal Models and Optimal Control
Goal: The primary goal of this thesis is to develop a multi-modal model which can provide cost functions to execute locomotion tasks for a legged robot.
- Investigating how available Multi-Modal Models can inform locomotion tasks.
- Implement fine-tuning the models to locomotion parameters or control targets.
- Integrate Optimal Control methods to realize locomotion policies informed by multi-modal models.
- Evaluation using simulation and robots available at DFKI-RIC.
Prior Knowledge:
- Background in Robotics, Computer Science, Mechatronics, Controls, or related fields.
- A good understanding of control theory as well as machine learning techniques. Experience with generative models and reinforcement learning is a plus.
- Experience in Programming with Python and C/C++
- Familiarity with hardware such as actuators and sensors is advantageous.
Related Work:
- Brohan, A., Chebotar, Y., Finn, C., Hausman, K., Herzog, A., Ho, D., Ibarz, J., Irpan, A., Jang, E., Julian, R. and Kalashnikov, D., 2023, March. Do as i can, not as i say: Grounding language in robotic affordances. In Conference on robot learning (pp. 287-318). PMLR.
- Liu, J., Liu, M., Wang, Z., An, P., Li, X., Zhou, K., Yang, S., Zhang, R., Guo, Y. and Zhang, S., 2024. Robomamba: Efficient vision-language-action model for robotic reasoning and manipulation. Advances in Neural Information Processing Systems, 37, pp.40085-40110.
For further information and application please contact Shubham Vyas at Shubham.Vyas[at]dfki.de.
Contact:
Deutsches Forschungszentrum für Künstliche Intelligenz GmbH
Robotics Innovation Center
Robert-Hooke-Str. 1
28359 Bremen, Germany
www.dfki.de/ric
Shubham Vyas
Phone: +49 421 17845 4157
shubham.vyas@dfki.de