|
|
| CARVIEW |
Deployment
We deploy the trained policy on the Unitree G1 in the real world and teleoperate it to perform various loco-manipulation tasks using our isomorphic exoskeleton hardware system.
1. Walking under changing upper-body poses.
2. Squatting under changing upper-body poses.
3. Squat to hold flower and transfer.
4. Squat to grasp a bottle.
5. Hand over and pick & place.
6. Step back and open oven.
7. Transfer a grasp from lower to higher.
8. Transfer a box from one shelf to another.
9. Push the man on a chair.
10. Hand over between two robots.
We further conduct some experiments to show the robustness of our policy.
1. Strong hitting.
2. Hit with a heavy ball.
To demonstrate the effectiveness of the isomorphic exoskeleton, we compare the task completion times across four different tasks between our hardware system and OpenTelevision.
1. Pick & Place.
2. Scan Barcode.
3. Hand Over.
4. Open Oven.
The completion times for these tasks are computed based on data from three different operators, with each operator performing the tasks three times. Our hardware system can accelerate the teleoperation by approximately 2 times, particularly in tasks that require radial movement.
Extensions
1. Simulation
We transfer the trained policies for Unitree G1 and Fourier GR-1 from Isaac Gym to scenes developed by GRUtopia, thus the robots can perform diverse loco-manipulation tasks more cost-effectively and in a wider range of scenarios than would be feasible in the real world.
1. Unitree G1 in GRUtopia.
2. Fourier GR-1 in GRUtopia.
3. Loco-Manipulation Task Completion in GRUtopia.
2. Imitation Learning
To validate the effectiveness of the demonstratons collected by HOMIE for IL algorithms, we design two distinct tasks, collect data by teleoperating, train with IL algorithm, and deploy in the real world. We achieve over 70% success rate, showing the feasibility of training IL with collected data.
1. Squat Pick.
2. Pick & Place.
Authors
@article{ben2024homie,
title={HOMIE: Humanoid Loco-Manipulation with Isomorphic Exoskeleton Cockpit},
author={Qingwei Ben, Feiyu Jia, Jia Zeng, Junting Dong, Dahua Lin, Jiangmiao Pang},
journal={arXiv preprint arXiv:2502.13013},
year={2025}
}If you have any questions, please contact Qingwei Ben. 🎉