The carla autoware bridge is now hosted and maintained here.

This repository contains a demonstrator of an autoware agent ready to be executed with CARLA.

The autoware agent is provided as a ROS package. All the configuration can be found inside the carla-autoware-agent folder.

The easiest way to run the agent is by building and running the provided docker image.

• Docker (19.03+)
• Nvidia docker (https://github.com/NVIDIA/nvidia-docker)

Firstly clone the carla autoware repository, where additional autoware contents are included as a submodule:

git clone --recurse-submodules https://github.com/carla-simulator/carla-autoware

Afterwards, build the image with the following command:

./build.sh

This will generate a carla-autoware:latest docker image.

1. Run a CARLA server.

You can either run the CARLA server in your host machine or within a container. To start a CARLA server within a docker container run the following command:

docker run -p 2000-2002:2000-2002 --runtime=nvidia --gpus all carlasim/carla:latest

You may find more information about running CARLA using docker here

2. Run the carla-autoware image:

./run.sh

This will start an interactive shell inside the container. To start the agent run the following command:

roslaunch carla_autoware_agent carla_autoware_agent.launch town:=Town01

The autoware-contents repository contains additional data required to run Autoware with CARLA, including the point cloud maps, vector maps and configuration files.