search

aCodeDog / legged-robots-manipulation

94 stars 10 forks source link

readme

legged-robots-manipulation

Introduction

legged-robots-manipulation is a loco-manipulation repository for (wheel-)legged robots. The code is built on legged_gym.

Project Page:wheel-legged-loco-manipulation (IROS Oral 2024)

The current repository contains airbot ,go2_arx,b2w_z1,aliengo_z1 and b2w. The current repository is a partial implementation of the paper.The repository is still under construction for various reasons, and will be releasing issac lab versions and vision-based versions in the near future.

To do list

airbot

loco-manipulation

go2_arx

loco-manipulation

aliengo_Z1

loco-manipulation

B2W_Z1

b2w

b2w is a locomotion demo for the Unitree b2w robot. locomotion

Installation

install isaac_gym,rsl_rl,legged_gym

  1. You are supposed to install the isaac_gym,rsl_rl and legged_gym.Please install the above dependencies (Isaac Gym Preview 4, rsl_rl, legged_gym) as described in legged_gym.

install legged-robots-manipulation

cd loco_manipulation_gym
pip install -e .

Train

python tran.py --task=b2w --rl_device=cuda:0

# or

python train.py --task=airbot  --rl_device=cuda:0

Citing

If you use this work, please cite this paper:

      @misc{wang2024armconstrained,
        title={Arm-Constrained Curriculum Learning for Loco-Manipulation of the Wheel-Legged Robot}, 
        author={Zifan Wang, Yufei Jia, Lu Shi, Haoyu Wang, Haizhou Zhao, Xueyang Li, Jinni Zhou, Jun Ma and Guyue Zhou},
        year={2024},
        eprint={2403.16535},
        archivePrefix={arXiv},
        primaryClass={cs.RO}
  }

IEEE concur reimbursement process reference:

IEEE_concur_reimbursement_process.pdf

Acknowledgements

This code builds upon following open-source code-bases. Please visit the URLs to see the respective LICENSES:

  1. https://github.com/leggedrobotics/legged_gym/tree/master
  2. https://github.com/aCodeDog/awesome-loco-manipulation
  3. https://github.com/MarkFzp/Deep-Whole-Body-Control