Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Learning recovery behavior in cluttered environments.
In this project, a student will explore RL-based control approaches for legged robots to enable our quadrupedal robot ANYmal to get up from the rough terrain with obstacles.
We aim to extend our previous work on the recovery controller [1] to the non-flat rough terrains with various obstacles. In this project, a student will explore RL-based control approaches for legged robots to enable our quadrupedal robot ANYmal to get up from the ground. To generate richier behaviors and encourage exploration during training, we would like to investigate various methods including hierarchical learning, kinematic motion primitives, privileged training, and so on.
[1] Lee, Joonho, Jemin Hwangbo, and Marco Hutter. "Robust recovery controller for a quadrupedal robot using deep reinforcement learning." arXiv preprint arXiv:1901.07517 (2019).
We aim to extend our previous work on the recovery controller [1] to the non-flat rough terrains with various obstacles. In this project, a student will explore RL-based control approaches for legged robots to enable our quadrupedal robot ANYmal to get up from the ground. To generate richier behaviors and encourage exploration during training, we would like to investigate various methods including hierarchical learning, kinematic motion primitives, privileged training, and so on.
[1] Lee, Joonho, Jemin Hwangbo, and Marco Hutter. "Robust recovery controller for a quadrupedal robot using deep reinforcement learning." arXiv preprint arXiv:1901.07517 (2019).
- Literature research
- Implementation of simulation environments
- Experiments with hardware
- Literature research - Implementation of simulation environments - Experiments with hardware
- Good problem-solving skills
- C++ & Python skills
- Experience with the machine learning tools (Pytorch)
- Must have taken courses related to machine learning
- Good problem-solving skills - C++ & Python skills - Experience with the machine learning tools (Pytorch) - Must have taken courses related to machine learning
Send your CV and transcript of records to Joonho Lee (jolee@ethz.ch)
Send your CV and transcript of records to Joonho Lee (jolee@ethz.ch)