Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Real robot deployment of model-based controller and state estimator (Unitree A1 robot)
In this project, you will implement one of the current state-of-the-art model-based controller or state estimation algorithm for legged robots in C++. We aim to deploy the new controller/estimator to Unitree A1 quadrupedal robot, and benchmark the performance.
- Semester project (14 weeks)
- Starting date: From July, 2021
- Prerequisite courses:
- Robot Dynamics or Computational Models of Motion for Character Animation and Robotics (required)
- Recursive Estimation (required)
- Model Predictive Control (optional)
- Required skills:
- C++ programming (required)
- experiences with ROS (optional)
In this project, you will implement one of the current state-of-the-art model-based controller or state estimation algorithm for legged robots in C++. We aim to deploy the new controller/estimator to Unitree A1 quadrupedal robot, and benchmark the performance.
Please see more details of this project at https://donghok.me/bio-inspired-robots/
Note. For students from a research or educational institute other than ETH
The Master's semester project and thesis project are part of the Master's degree program at ETH Zurich and we cannot provide any financial support for cost of living and neither legal support for your visa process. If you are interested in this project however, there are plenty of opportunities to visit ETH for research exchanging purposes. So please make sure you can visit ETH and Switzerland with legitimated permission supported by one of these programmes before you contact us.
- Semester project (14 weeks) - Starting date: From July, 2021 - Prerequisite courses: - Robot Dynamics or Computational Models of Motion for Character Animation and Robotics (required) - Recursive Estimation (required) - Model Predictive Control (optional) - Required skills: - C++ programming (required) - experiences with ROS (optional)
In this project, you will implement one of the current state-of-the-art model-based controller or state estimation algorithm for legged robots in C++. We aim to deploy the new controller/estimator to Unitree A1 quadrupedal robot, and benchmark the performance.
Please see more details of this project at https://donghok.me/bio-inspired-robots/
Note. For students from a research or educational institute other than ETH
The Master's semester project and thesis project are part of the Master's degree program at ETH Zurich and we cannot provide any financial support for cost of living and neither legal support for your visa process. If you are interested in this project however, there are plenty of opportunities to visit ETH for research exchanging purposes. So please make sure you can visit ETH and Switzerland with legitimated permission supported by one of these programmes before you contact us.
At the end of this project, you are supposed to deliver
- C++ software running on onboard PC of A1 robot
- C++ estimator/controller implementation
- Benchmark results (comparison with our current control/estimation stacks)
- Documentation (report or wiki style docs)
At the end of this project, you are supposed to deliver
- C++ software running on onboard PC of A1 robot - C++ estimator/controller implementation - Benchmark results (comparison with our current control/estimation stacks) - Documentation (report or wiki style docs)
Please contact Dongho Kang or Computational Robotics Lab:
kangd@ethz.ch / crl-jobs@inf.ethz.ch
Please contact Dongho Kang or Computational Robotics Lab: