This is a demostration of the integration of AutowareAI and the CARLA simulator.
- This demo works on most Linux distributions with CUDA installed (you need an NVIDIA graphics card).
- Install NVIDIA drivers and verify the drivers with:
nvidia-smi - This demo requires Docker, NVIDIA Container Toolkit, and ADE to be installed.
- For a local installation of the requirements run:
bash install_dependencies.shafter you have cloned this repository. (see Getting Started for more information) - To remove the installed binaries and the docker images run:
bash remove_dependencies.sh
- Create your workspace directory (you can choose your own name or use an existing folder):
mkdir adehome - Go to your ADE home directory:
cd adehome - Create an empty .adehome file:
touch .adehome - Clone this repository:
git clone https://github.com/eLearningHub/AutowareAI-CARLA.git - Enter the clone directory:
cd AutowareAI-CARLA - Here you can choose to install the dependencies with the
bash install_dependencies.sh(requiresmakepackage and root privileges) - Run ADE:
ade start --enter - Inside ADE, go to the clone directory:
cd AutowareAI-CARLA - Inside ADE, run Carla:
source run-carla.sh - Open another terminal at the ADE home directory and run:
ade enter - Inside ADE, run:
cd AutowareAI-CARLA; source launch-autoware.sh - Press P for the vehicle to enter into the autonomous mode. Voila!