One repo to rule them all
git clone --recurse-submodules https://github.com/cornellev/rc-brain.git
cd rc-brain && sudo ./install.sh
docker build -t cev-rc-autonomy .
Then run either
docker run --rm ros2-autonomy launch
or
docker run --rm ros2-autonomy teleop
Run ./scripts/deploy/arduino_setup.sh the first time arduino is setup.
Run ./scripts/deploy/arduino_deploy.sh to compile and flash code to the arduino.
If you're not using the DevContainer, please make sure to install clang-format, ideally version 19! This will help keep the formatting of our C++ code formatted consistently. Python code is formatted by black. The DevContainer will install the Black extension for you, which ships with black, but otherwise you'll need to install it in VSCode or whatever other IDE you're using.
If you're just developing, building Docker images will probably take too long. You can deploy normally just by running scripts/deploy.sh. It will drop you into an interactive session at the end so you can launch whatever you want to.
Our Docker deployment pipeline works by pushing Docker images to a local registry and pulling them down on the mini cars. This avoids both long build times on the mini cars and the long time it takes to transfer whole images (as we can take advantage of Docker's cache).
You'll have to do a bit of setup to get your computer to use the local registry. Note that you must be connected to cev-router.
Add the following configuration to /etc/docker/daemon.json:
{
"insecure-registries" : [ "mini-server:5000" ]
}Then, restart the docker daemon (with systemctl restart docker on linux).
If you have errors regarding buildx or buildkit, please make sure you have Docker BuildX installed.
If there are still errors regarding some "Exec format error", then run
docker run --privileged --rm tonistiigi/binfmt --install all
which will install qemu for cross-platform compilation.
To send and receive data over ros while on the same network, simply run
export ROS_DOMAIN_ID=0
export ROS_LOCALHOST_ONLY=0
on all devices.