Setup process and ROS code for using Unitree GO2 EDU quadruped with essential ROS functionality. Functionalities include pointcloud and laserscan acquisition, SLAM, Nav2 navigation, camera data, etc.

This guide assumes you have a GO2 or other robot equipped with an Intel Realsense D435i and a Livox LiDar

and that both of these components have been mounted on the robot following the official docs.

It also assumes the robot is running ROS2 Foxy, but this shouldn't cause many issues.

Go2 Expansion Board Networking Setup

This section outlines how to get the expansion board component of the Go2 dog (an NVIDIA Orin board) connected to an external Wi-Fi network and how to access the ROS topics of the Go2's internal board on the expansion board. This is a separate process than connecting the Go2 to Wi-Fi as outlined in the official Unitree docs. Before starting this section, you will need:

1. A monitor, keyboard, and mouse
2. A BrosTrend AC1L USB Wi-Fi adapter (or equivalent Ubuntu-compatible Wi-Fi adapter)
3. A Type-C USB hub with an Ethernet port
4. An HDMI cable that can be plugged into the aforementioned USB hub
5. An Ubuntu host PC with Ethernet port or equivalent

Step 1:

Follow the quick start section of the Unitree docs as found in the Unitree Docs and the app binding section. Ensure you connect to the dog through the app through the Wi-Fi mode. Once connected through the app, record the Go2's IP (Device/Data/Automatic Machine Inspection) as it will be needed later.

It's also a good idea to go ahead and do the Go2 ROS2 setup that matches with your distro.

Step 2:

Plug the USB hub into the Type-C port on the back of the Go2 expansion board. Plug the HDMI into the USB hub, and into the monitor. Connect the keyboard and mouse to the USB hub. The monitor should display the Ubuntu GUI. Login to the expansion board (default password is 123).

Go to the network settings. Disable the PCI Ethernet connection. This will allow the board to connect to the Internet. Plug the Ethernet cable into the USB hub, and then into an Internet-connected USB port. Check if the expansion board can connect to the Internet, e.g. by ping google.com.

Step 3:

Download the BrosTrend AC1L adapter Linux drivers (or the equivalent for your Wi-Fi adapter) by running

sudo sh -c 'wget linux.brostrend.com/install -O /tmp/install && sh /tmp/install'

Follow the instructions in the terminal. Ensure all drivers are installed correctly.

Step 4:

Once the drivers have been installed, go back to network settings, and connect to your Wi-Fi network of choice. Unplug the Ethernet cable from the USB hub, and ensure you can access the Internet on the Wi-Fi connection. Record the IP of the expansion board on the Wi-Fi so you can acces it later. If there are no problems, re-enable the PCI Ethernet connection; this will resume the communication between the expansion board and the Go2's internal board.

In network settings, open the wired connection settings and change the DHCP settings IPv4 address to 192.168.123.99 and make sure the mask is 255.255.255.0. This will allow you to access the topics published by the Go2's internal board on the expansion board.

Step 5:

Run the commands echo 'net.core.rmem_max=52428800' | sudo tee -a /etc/sysctl.conf and sudo sysctl -p. This will increase the expansion board's maximum socket buffer size. To utilize this with ROS, go to ~/cyclonedds_ws/cyclonedds.xml and add the following lines if they are not there already:

<Internal>
    <MinimumSocketReceiveBufferSize>50MB</MinimumSocketReceiveBufferSize>
</Internal>

This will allow the expansion board to handle more data across its Ethernet bus.

Finally, increase the power level of the expansion board to 25W or higher.

Go2 Expansion Board Intel RealSense Setup

The Go2's expansion board does not come preconfigured to use the D435i depth camera with Python bindings. Since the board uses an ARM architecture, the bindings can't be installed directly through pip either, and must be built from source. Steps 1-4 outlined below are taken directly from this official process from a RealSense Git issue. Steps 5-6 were added by me. The process should be nearly identical for any ARM system, and should be even simpler if your target platform does not have an ARM architecture.

Step 1:

The first step in this process is to download the source code of librealsense from the Releases page as a zip file. The source code zip file can always be found in the "Assets" list at the bottom of the information listing for each SDK version on the Releases page.

https://github.com/IntelRealSense/librealsense/releases/

Step 2:

After downloading the zip file, its contents should be extracted so that you have a librealsense folder.At this point you can use the RSUSB installation method to install librealsense without dependence on Linux versions or kernel versions and without the need for patching. This makes this installation method particularly suited to Arm devices such as Jetson.

The backend installation method requires an internet connection. The steps are:

Go to the librealsense root directory. Create a folder within it called build and then go to this new folder using this instruction: mkdir build && cd build

Step 3:

Now that you are in the build directory, run the CMake build instruction below to install librealsense and the Python bindings over the internet connection: cmake ../ -DFORCE_RSUSB_BACKEND=ON -DBUILD_PYTHON_BINDINGS:bool=true -DPYTHON_EXECUTABLE=...

The above statement is a basic one that should test whether the build is likely to succeed or not. (Replace the dots with the path to your Python exec)

Step 4:

If it does succeed then you can try a more advanced build, which builds the example programs and includes optimizations such as building with CUDA support for faster alignment processing on devices such as Jetson that are equipped with an Nvidia graphics GPU.

cmake ../-DFORCE_RSUSB_BACKEND=ON -DBUILD_PYTHON_BINDINGS:bool=true -DPYTHON_EXECUTABLE=... -DCMAKE_BUILD_TYPE=release -DBUILD_EXAMPLES=false -DBUILD_GRAPHICAL_EXAMPLES=false -DBUILD_WITH_CUDA:bool=true

(The original issue has this command running with examples and graphical examples on. This can cause issues with the following step, so they have been disabled here.)

Step 5:

Run make -j$(nproc)

Step 6:

If everything builds correctly, run find . -name "pyrealsense2*.so" within the build directory to file the .so filepath. Once you have this path, find the site packages location with python3 -m site --user-site. Then, use cp to copy the .so to the site packages location.

Livox Setup

Most of this process is taken care of by mounting the Livox according to the official docs from Unitree or whatever the target platform is. The interfacing of the Livox with the robot should be taken care of by the packages in this repo. However, you may need to change the ip parameter within the src/CPSL_ROS_livox_ros_driver2/config/HAP_config.json file to the IP of your Livox device (IP can be found on a sticker on the physical device).

Go2 Internal IP Configuration

In order for the Go2 to receive commands from packages like Nav2, cmd_vel messages must be translated and sent to the Go2's internal board which handles all of its movement. As such, the CmdVelTranslator.py node within the dog_utilities needs a parameter internal_board_ip passed to it either when calling the main launch file, or when running the node itself.

Using this repo

1. Clone this repo and build and source the workspace
2. Open 3 terminal windows and run the following
   • ros2 launch launcher dog.launch.py internal_board_ip:=YOUR_IP_HERE collect_realsense:=true or false
   • ros2 launch cpsl_ros2_sensors_bringup ugv_sensor_bringup.launch.py
   • ros2 launch cpsl_nav slam.launch.py scan_topic:=/livox/scan_best_effort
3. For getting a functional transform tree and mapping, that's all you need. If you want to issue Nav2 commands, open another terminal and run
   • ros2 launch cpsl_nav nav2_archived.py scan_topic:=/livox/scan_best_effort