This repository contains a Python implementation of the WebRTC driver to connect to the Unitree Go2 Robot. WebRTC is used by the Unitree Go APP and provides high-level control through it. Therefore, no jailbreak or firmware manipulation is required. It works out of the box for Go2 AIR/PRO/EDU models.
The currently supported firmware packages are:
- 1.1.1
- 1.0.19 - 1.0.25
There are video (recvonly) and audio (sendrecv) channels in WebRTC that you can connect to. Check out the examples in the /example
folder.
There is a lidar decoder built in, so you can handle decoded PoinClouds directly. Check out the examples in the /example
folder.
The driver supports three types of connection methods:
-
AP Mode: Go2 is in AP mode, and the WebRTC client is connected directly to it:
Go2WebRTCConnection(WebRTCConnectionMethod.LocalAP)
-
STA-L Mode: Go2 and the WebRTC client are on the same local network. An IP or Serial number is required:
Go2WebRTCConnection(WebRTCConnectionMethod.LocalSTA, ip="192.168.8.181")
If the IP is unknown, you can specify only the serial number, and the driver will try to find the IP using the special Multicast discovery feature available on Go2:
Go2WebRTCConnection(WebRTCConnectionMethod.LocalSTA, serialNumber="B42D2000XXXXXXXX")
-
STA-T mode: Remote connection through remote Unitrees TURN server. Could control your Go2 even being on the diffrent network. Requires username and pass from Unitree account
Go2WebRTCConnection(WebRTCConnectionMethod.Remote, serialNumber="B42D2000XXXXXXXX", username="[email protected]", password="pass")
The driver has a built-in Multicast scanner to find the Unitree Go2 on the local network and connect using only the serial number.
cd ~
sudo apt update
sudo apt install python3-pip
sudo apt install portaudio19-dev
git clone --recurse-submodules https://github.com/legion1581/go2_webrtc_connect.git
cd go2_webrtc_connect
pip install -e .
Example programs are located in the /example directory.
A big thank you to TheRoboVerse community! Visit us at TheRoboVerse for more information and support.
Special thanks to the tfoldi WebRTC project and abizovnuralem for adding LiDAR support!
If you like this project, please consider buying me a coffee: