This branch contains the ROS+Gazebo simulation for the Hector humanoid robot.
For humanoid ROS model with arms use the ROS_Humanoid_Simulation branch.
Introduction Video: https://youtu.be/NcW-NFwjMh0
Video associated with preprint "Dynamic Loco-manipulation on HECTOR: Humanoid for Enhanced ConTrol and Open-source Research": https://youtu.be/-r0QoxQgshk
- Boost (version 1.5.4 or higher)
- CMake (version 2.8.3 or higher)
- LCM (version 1.4.0 or higher)
- ROS Neotic
- Gazebo 11
- Eigen3 (>3.3)
- unitree_legged_sdk
- qpOASES
- ROS_Packages
sudo apt-get install ros-noetic-controller-manager ros-noetic-ros-control ros-noetic-ros-controllers ros-noetic-joint-state-controller ros-noetic-effort-controllers ros-noetic-velocity-controllers ros-noetic-position-controllers ros-noetic-robot-controllers ros-noetic-robot-state-publisher ros-noetic-gazebo-ros-pkgs ros-noetic-gazebo-ros-control
If you want to simulate with Gazebo, we recommend x86 platform. ARM platform is not suggested for simulation. So, if you run this code on ARM platform, please remove Simulation related folder first.
The current system environment is:
- Ubuntu 20.04 + ROS Noetic* (recommended, tested stable)
Use command to open .bashrc file:
gedit ~/.bashrc
Make sure the following exist in your ~/.bashrc
file or export them in terminal. noetic
, gazebo-11
and ~/catkin_ws
should be replaced in your own case.
source /opt/ros/noetic/setup.bash
source /usr/share/gazebo-11/setup.sh
source ~/catkin_ws/devel/setup.bash
export ROS_PACKAGE_PATH=~/catkin_ws:${ROS_PACKAGE_PATH}
export GAZEBO_PLUGIN_PATH=~/catkin_ws/devel/lib:${GAZEBO_PLUGIN_PATH}
export LD_LIBRARY_PATH=~/catkin_ws/devel/lib:${LD_LIBRARY_PATH}
cd ~/catkin_ws
catkin_make
NOTE: If it is the first time to compile, Please compile the laikago_msgs first by following command:
complie the package
catkin_make -DCMAKE_BUILD_TYPE=Release
-
roslaunch unitree_gazebo biped.launch
The robot should be standing on the ground -
In a new terminal, enter and source your workspace, then run:
rosrun hector_control hector_ctrl
-
Click the start button at the bottom of the simulator, the robot should stand up/move away
-
In some occasions the controller does not kick in after starting, please terminate the controller with ctrl + \. Then go back to the simulator, pause, and reset (ctrl + R). Rerun controller.
- Inside the terminal window, use W or S to control x direction speed
- Use A or D to control robot turning
- Use J or L to control y direction speed
Thank you for choosing our software for your research and development, we highly appreciate your citing our work:
- HECTOR Project: "Dynamic Loco-manipulation on HECTOR: Humanoid for Enhanced ConTrol and Open-source Research" https://arxiv.org/pdf/2312.11868.pdf
@article{li2023dynamic,
title={Dynamic Loco-manipulation on HECTOR: Humanoid for Enhanced ConTrol and Open-source Research},
author={Li, Junheng and Ma, Junchao and Kolt, Omar and Shah, Manas and Nguyen, Quan},
journal={arXiv preprint arXiv:2312.11868},
year={2023}
}
- Force-and-moment-based Locomotion MPC: "Force-and-moment-based model predictive control for achieving highly dynamic locomotion on bipedal robots" https://arxiv.org/abs/2104.00065
@inproceedings{li2021force,
title={Force-and-moment-based model predictive control for achieving highly dynamic locomotion on bipedal robots},
author={Li, Junheng and Nguyen, Quan},
booktitle={2021 60th IEEE Conference on Decision and Control (CDC)},
pages={1024--1030},
year={2021},
organization={IEEE}
}
Junheng Li -- junhengl@usc.edu Yiyu Chen -- yiyuc@usc.edu
Please read the License.md for details.
The authors would like to express special thanks to MIT Biomimetic Lab for providing the cheetah MPC framework and Unitree Robotics for providing the Unitree gazebo simulation framework.