Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Robustify Whole-Body MPC for Legged Mobile Manipulator
The goal of this project is to increase the locomotion capabilities of the whole-body Model predictive control (MPC) planner for the quadrupedal manipulator, ALMA, by integrating the perceptive MPC locomotion. This way, the robot can execute manipulation tasks while traversing rough terrain, e.g., slopes and stairs.
A unified MPC framework for the locomotion and manipulation of ALMA has been introduced, allowing whole-body dynamic motions, such as object throwing, collision avoidance, and door opening [1][2]. However, the framework does not perceive the terrain and, thus, assumes an even terrain. On the other hand, the locomotion capabilities of MPC have been greatly boosted by motion planning based on the foothold constraints extracted from an elevation map [3]. Together with the loop-shaping technique reducing the touchdown impact [4][5], ANYmal can agilely locomote over various terrain. Hence, in this project, we will combine the two state-of-the-art MPC planners to tackle a broader range of real-world challenges with ALMA.
[1] Sleiman, Jean-Pierre, et al. "A unified mpc framework for whole-body dynamic locomotion and manipulation." IEEE Robotics and Automation Letters 6.3 (2021): 4688-4695.
[2] Chiu, Jia-Ruei, et al. "A collision-free mpc for whole-body dynamic locomotion and manipulation." 2022 International Conference on Robotics and Automation (ICRA). IEEE, 2022.
[3] Grandia, Ruben, et al. "Perceptive locomotion through nonlinear model predictive control." arXiv preprint arXiv:2208.08373 (2022).
[4] Grandia, Ruben, et al. "Frequency-aware model predictive control." IEEE Robotics and Automation Letters 4.2 (2019): 1517-1524.
[5] Grandia, Ruben, et al. "Feedback mpc for torque-controlled legged robots." 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2019.
A unified MPC framework for the locomotion and manipulation of ALMA has been introduced, allowing whole-body dynamic motions, such as object throwing, collision avoidance, and door opening [1][2]. However, the framework does not perceive the terrain and, thus, assumes an even terrain. On the other hand, the locomotion capabilities of MPC have been greatly boosted by motion planning based on the foothold constraints extracted from an elevation map [3]. Together with the loop-shaping technique reducing the touchdown impact [4][5], ANYmal can agilely locomote over various terrain. Hence, in this project, we will combine the two state-of-the-art MPC planners to tackle a broader range of real-world challenges with ALMA.
[1] Sleiman, Jean-Pierre, et al. "A unified mpc framework for whole-body dynamic locomotion and manipulation." IEEE Robotics and Automation Letters 6.3 (2021): 4688-4695.
[2] Chiu, Jia-Ruei, et al. "A collision-free mpc for whole-body dynamic locomotion and manipulation." 2022 International Conference on Robotics and Automation (ICRA). IEEE, 2022.
[3] Grandia, Ruben, et al. "Perceptive locomotion through nonlinear model predictive control." arXiv preprint arXiv:2208.08373 (2022).
[4] Grandia, Ruben, et al. "Frequency-aware model predictive control." IEEE Robotics and Automation Letters 4.2 (2019): 1517-1524.
[5] Grandia, Ruben, et al. "Feedback mpc for torque-controlled legged robots." 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2019.
- Literature & OCS2 toolbox review
- Integrate gait adaption, loop shaping technique, and terrain awareness into ALMA MPC.
- Simulation and hardware tests
- Literature & OCS2 toolbox review - Integrate gait adaption, loop shaping technique, and terrain awareness into ALMA MPC. - Simulation and hardware tests
- Excellent knowledge in C++ & ROS
- Knowledge of MPC/optimal control and robot dynamics
- Autonomous working style.
- Excellent knowledge in C++ & ROS - Knowledge of MPC/optimal control and robot dynamics - Autonomous working style.
Jia-Ruei Chiu (jichiu@student.ethz.ch) Please include your CV and up-to-date transcript.
Jia-Ruei Chiu (jichiu@student.ethz.ch) Please include your CV and up-to-date transcript.