Introduction

This repository contains USC GVL's custom Humanoid data collection pipeline, along with the re-engineering of Unitree's teleoperation of the G1 humanoid robots. For teleoperating and collecting data using the H1, please check out the H1 branch.

Environment Requirements

We use micromamba instead of conda for its lightweight.

Inverse Kinematics Libraries (Host only)

micromamba create -n tv python=3.8
micromamba activate tv
micromamba install pinocchio -c conda-forge
pip install meshcat
pip install casadi

Install Unitree SDK2 Python (Host only)

git clone https://github.com/unitreerobotics/unitree_sdk2_python.git
cd unitree_sdk2_python
pip install -e .

Note

You need to fix the dds issue by manually removing any code relevant to __domain__

TeleVision and Apple Vision Pro configuration (Host only)

pip install -r requirements.txt

Apple Vision Pro (AVP) Setup Local streaming (Host and AVP)

Apple does not allow WebXR on non-https connections. To test the application locally, we need to create a self-signed certificate and install it on the client. You need a ubuntu machine and a router. Connect the VisionPro and the ubuntu machine to the same router.

1. install mkcert: https://github.com/FiloSottile/mkcert
2. check local ip address:

ifconfig | grep inet

Suppose the local ip address of the ubuntu machine is 192.168.123.2

3. create certificate:

mkcert -install && mkcert -cert-file cert.pem -key-file key.pem 192.168.123.2 localhost 127.0.0.1

ps. place the generated cert.pem and key.pem files in teleop.

cp cert.pem key.pem teleop/

4. open firewall on server:

sudo ufw allow 8012

5. install ca-certificates on VisionPro:

mkcert -CAROOT

Copy the rootCA.pem via AirDrop to VisionPro and install it.

Settings > General > About > Certificate Trust Settings. Under "Enable full trust for root certificates", turn on trust for the certificate.

settings > Apps > Safari > Advanced > Feature Flags > Enable WebXR Related Features

6. open the browser on Safari on VisionPro and go to https://192.168.123.2:8012?ws=wss://192.168.123.2:8012

7. Click Enter VR and Allow to start the VR session.

Image Server environment (Robot only)

On G1's PC, install the image server's requirements

micromamba create -n vision python=3.8
micromamba activate vision
pip install pyrealsense2 opencv-python zmq numpy

Lidar (Host only)

Compile and lidar's C++ point cloud recorder program

mkdir Livox-SDK2/build && cd Livox-SDK2/build
cmake ..
make -j
mv point_cloud_recordedr/point_cloud_recorder ../../avp_teleop/teleop

Start teleoperating and data collection

Image Server

Copy realsense_server.py in the teleop/image_server directory to the PC of Unitree G_1, and execute the following command in the PC:

micromamba activate vision
python realsense_server.py

Data collection metadata preparation

1. Write down the task information in a json file inside teleop/task_defs directory
2. In the teleop/ directory, run python taskcreator.py to generate the task metadata. Take a look at teleop/task_defs/example.json for an example task's json format.

Teleoperating Data Collection Instruction

Warning : All persons must maintain an adequate safety distance from the robot to avoid danger!

1. Connect your host computer to G1. Then, set up your local IP address to 192.168.123.123 with netmask 255.255.255.0 on G1's network interface.
2. Connect both your computer and AVP to the same local router wifi in which you set up your cert with.
3. Open robot and set to sports mode (using remote control by pressing L1 + A, then L1 + UP, and lastly R1 + X when the G1 is gently touching the ground).
4. On G1 PC: start the image server as above.
5. On host computer: run

export CYCLONEDDS_URI="<CycloneDDS><Domain><General><NetworkInterfaceAddress>192.168.123.123</NetworkInterfaceAddress></General></Domain></CycloneDDS>"

6. On host computer: run python main.py --robot g1 under the teleop/ directory and wait until the robot is in ready state. The terminal should signal both "master" and "worker" processes are waiting for starting signal.
7. On AVP, connect to robot using https://<host_ip_address_on_your_local_router>:8012/?ws=wss://<host_ip_address_on_your_local_router>:8012. Then, press Enter VR and then Allow to enter the web interface for teleoperating the G1.
8. Back on host computer, enter s to start recording an episode.
9. Type q and enter if the episode is successful, otherwise d and enter to discard the last session.
10. Repeat by pressing s to start recording the next episode. Record 40 episodes for each task.
11. Type exit to shut down the program.

Acknowledgement

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

1. https://github.com/unitreerobotics/avp_teleoperate
2. https://github.com/OpenTeleVision/TeleVision
3. https://github.com/dexsuite/dex-retargeting
4. https://github.com/vuer-ai/vuer
5. https://github.com/stack-of-tasks/pinocchio
6. https://github.com/casadi/casadi
7. https://github.com/meshcat-dev/meshcat-python
8. https://github.com/zeromq/pyzmq
9. https://github.com/unitreerobotics/unitree_dds_wrapper
10. https://github.com/tonyzhaozh/act
11. https://github.com/facebookresearch/detr