This project is currently organized using submodules. In more stable releases, it will be managed with a package manager like pip. For now in an alpha stage of development, we stay light on our feet by using submodules and poetry

NOTE: This currently only works on a Linux distribution (tested on Ubuntu 20.04 and 22.04). It also only works with Python 3.10 (to be expanded in the future).

While Carla does not require use of a GPU, it is highly recommended. Using libraries like mmdet (for perception, as in avstack-core) will require a gpu.

First, clone the repository and submodules.

git clone --recurse-submodules https://github.com/avstack-lab/carla-sandbox.git
git submodule update --recursive

Dependencies are managed with poetry. This uses the pyproject.toml file to create a poetry.lock file. To install poetry, see this page.

If docker is not installed, install with something like

curl https://get.docker.com | sh \
  && sudo systemctl --now enable docker

Verify docker with docker run hello-world.

If nvidia-docker2 is not installed, run something like

distribution=$(. /etc/os-release;echo $ID$VERSION_ID) \
      && curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg \
      && curl -s -L https://nvidia.github.io/libnvidia-container/$distribution/libnvidia-container.list | \
            sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' | \
            sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list
sudo apt-get update
sudo apt-get install -y nvidia-docker2
sudo systemctl restart docker

Verify nvidia-docker2 is working with

sudo docker run --rm --gpus all nvidia/cuda:11.0.3-base-ubuntu20.04 nvidia-smi

Will take about 2-3 minutes, depending on your cpu count.

poetry install

Will take about 5 minutes the first time you download.

./initialize.sh

For example, run

cd run_carla
./run_carla_0913.sh

The examples are run as shell scripts. Therefore, you have to first activate the environment with

poetry shell

Try running a non-perception example such as:

cd examples
./run_a0_autopilot.sh

Note: You should be in the examples folder before running the example scripts.

With the carla server already running, run the dataset collection script:

poetry shell
cd examples
./run_a1_dataset_collection.sh

This will log data to the folder sim_results/. To use the dataset, we need to run postprocessing. With the poetry shell activated, navigate to submodules/lib-avstack-carla. Run the postprocessing:

python postprocess_carla_objects.py ../../examples/sim_results

This will take some time, depending on how long you ran the simulation. After postprocessing is complete, inspect the dataset by going back to the notebooks/ folder and opening test_carla_dataset.ipynb.

See CONTRIBUTING.md for further details.

Copyright 2023 Spencer Hallyburton

AVstack specific code is distributed under the MIT License.