This is the monorepo for all Tartan-AUV vehicle code, tools, and infrastructure.

Follow these instructions to set up the development environment and simulator on a fresh machine.

• Download VMware Fusion 13 Go to VMware.com and download Fusion and Workstation in the VMWare products. You will need to create a broadcom account. Download VMware fusion 13
• Create a Virtual Machine: Install the newest version of Ubuntu.
  • Disk Space: 30GB minimum (40GB recommended).
  • Click the wrench icon in VMware fusion and increase size, then run

    sudo lvextend -l +100%FREE /dev/ubuntu-vg/ubuntu-lv
    sudo resize2fs /dev/ubuntu-vg/ubuntu-lv
    

• Enable OpenSSH: Ensure you can access the VM remotely.
• Install GUI: Install XFCE (a lightweight GUI) to display the simulator later:

  sudo apt update
  sudo apt install xfce4

• Connect via SSH:
  1. Find the IP address of the VM:

     ip a

  2. Use this IP address to SSH into the machine for the remaining steps.

Docker Engine

1. Install the Docker Engine following the official guide: Install Docker on Ubuntu.
2. (Optional) Add your user to the docker group to run docker without sudo:

   sudo usermod -aG docker $USER

3. Authenticate with the GitHub Container Registry:

   docker login ghcr.io

   • Username is your github username, password is your github key
   • Settings -> developer settings -> personal access tokens -> tokens (classic).
   • Turn on repo, write:packages, user, delete:packages, admin:ssh_signing_key

Git Configuration

1. Configure global Git settings:

   git config --global user.name "Your Name"
   git config --global user.email "your.email@example.com"

2. Set up your SSH key with GitHub to allow for authentication during cloning.

1. Clone the repository:

   git clone --recurse-submodules https://github.com/Tartan-AUV/tauv-mono.git

2. Build the Docker container:

   cd tauv-mono/containers
   ./build_desktop_docker.sh
   docker compose up -d tauv-desktop

3. Attach to the container: Using the Dev Containers extension on VSCode, you should be able to attach to the container. Dev Containers: Attach to Running Container

   • Your VSCode may complain about unsafe repositories. This is simply because our code is segmented into submodules, and you can just mark everything as safe.
   • Can also use docker exec -it [Container Name] bash to attach

4. Build the ROS 2 Workspace: Once inside the container, build the workspace and source the setup script make sure you are in ros_ws folder :

   colcon build --symlink-install
   source install/setup.bash

   • Recommended 8GB of memory.
   • If it freezes while building it might have run out of memory, run this:

     colcon build --symlink-install --parallel-workers 1
     

     The --parallel-workers flag restricts building to one package at a time. The development environment has now been built.

ros2 launch tauv_sim desktop_sim.launch.py

Stonefish opens a GUI window — X11 access must be granted first. Add this to your ~/.bashrc so it runs on every login:

xhost +local:

If your machine has an NVIDIA GPU, the simulator will crash/lag without GPU passthrough. Two changes are needed:

1. In containers/docker-compose.yaml, under tauv-desktop, add:

   runtime: nvidia
   environment:
     - NVIDIA_VISIBLE_DEVICES=all
     - NVIDIA_DRIVER_CAPABILITIES=all

   These lines are already present on green's machine and marked with # NOTE (green) comments — do not commit them for general use.

2. Install the NVIDIA Container Toolkit on the host if not already present:

   sudo apt install nvidia-container-toolkit
   sudo nvidia-ctk runtime configure --runtime=docker
   sudo systemctl restart docker

Docker

• To be added

ROS

• To be added

• SSH Via Tailscale or connecting to router direclty(we only have 3 seats total on tailnet unfortunately)

1. Ensure you are in tauv-mono/containers and run ./build_osprey_docker.sh
   • only needs to be done if a modification has been made to the dockerfiles
2. Run The container
   • If it is not running already `docker compose up -d osprey-orin-user'
   • Attach to it by docker exec -it [Container name] bash
3. ROS Stuff
   • Will be in the ros_ws folder

Each package will contain a more detailed readme in their folder