Autoware Install

Ubuntu 22.04

ROS2 humble

nvidia-smi: Driver Version: 555.42.02 CUDA Version: 12.5

Prepare

make sure there is no old 'autoware', 'autoware_data', 'autoware_map' folder

cd ~

update and git install

sudo apt-get -y update
sudo apt-get -y install git

git clone

git clone https://github.com/autowarefoundation/autoware.git
gdown -O ~/autoware_map/ 'https://docs.google.com/uc?export=download&id=1499_nsbUbIeturZaDj7jhUownh5fvXHd'
unzip -d ~/autoware_map ~/autoware_map/sample-map-planning.zip

Memory check (In another terminal)

free -h

if you do not have 32G memory, do the following:

Remove the current swapfile

sudo swapoff /swapfile
sudo rm /swapfile

Create a new swapfile

sudo fallocate -l 32G /swapfile
sudo chmod 600 /swapfile
sudo mkswap /swapfile
sudo swapon /swapfile

Check if the change is reflected

free -h

Option-1(Source-Build): Workspace Setup

In autoware folder

Environment Setup (first time)

cd autoware
./setup-dev-env.sh

Build workspace

mkdir src
vcs import src < autoware.repos
sudo apt update
rosdep update
rosdep install -y --from-paths src --ignore-src --rosdistro $ROS_DISTRO
colcon build --symlink-install --cmake-args -DCMAKE_BUILD_TYPE=Release

colon build will take up to 1 hour, could quit and terminal all other processes to avoid stuck.

Option-2 (Docker-Build): Development setup

Docker Image: ghcr.io/autowarefoundation/autoware:latest-devel-cuda In autoware folder

Environment Setup (first time)

cd autoware
./setup-dev-env.sh -y --download-artifacts docker

Pull docker

./docker/run.sh --devel --map-path ~/autoware_map/sample-map-planning --workspace ~/autoware

After these, you should launch Autoware and rviz successfully, and you could see the same screen as shown in Planning simulation

Local folder 'autoware' will be mounted on '/workspace' by default.

Option-3 (Docker-Build) OpenADKit Install

Docker Images: ghcr.io/autowarefoundation/autoware-universe:latest-cuda

git clone https://github.com/autowarefoundation/autoware.git
cd autoware

Environment Setup (first time)

cd autoware
./setup-dev-env.sh -y --download-artifacts docker

Option-3.1 (Build-Own)

Login Docker

rocker --nvidia --x11 --user --volume $HOME/autoware_data -- ghcr.io/autowarefoundation/autoware-universe:latest-cuda

This will mount your local folder $~/autoware_data$ to your container and connect your local x11 rviz visualization to this container.

Docker prepare

cd ~
sudo apt-get -y update
sudo apt-get -y install git
sudo apt-get install unzip
git clone https://github.com/autowarefoundation/autoware.git
gdown -O ~/autoware_map/ 'https://docs.google.com/uc?export=download&id=1499_nsbUbIeturZaDj7jhUownh5fvXHd'
unzip -d ~/autoware_map ~/autoware_map/sample-map-planning.zip

Workspace setup

cd autoware
mkdir src
vcs import src < autoware.repos
sudo apt update
rosdep update
rosdep install -y --from-paths src --ignore-src --rosdistro $ROS_DISTRO
colcon build --symlink-install --cmake-args -DCMAKE_BUILD_TYPE=Release

Option-3.2 (Use-Pre-Build)

On your local machine

docker pull ariiees/openadk:build

You can check this imageID using the following command:

docker images

Put this imageID into the following command:

docker run -it  -v $HOME/autoware_data:$HOME/autoware_data  -e DISPLAY -e TERM   -e QT_X11_NO_MITSHM=1   -e XAUTHORITY=/tmp/.docker5guv3arw.xauth -v /tmp/.docker5guv3arw.xauth:/tmp/.docker5guv3arw.xauth   -v /tmp/.X11-unix:/tmp/.X11-unix   -v /etc/localtime:/etc/localtime:ro [imageID] /bin/bash

Inside the docker, test tutorial

source ~/autoware/install/setup.bash
ros2 launch autoware_launch planning_simulator.launch.xml map_path:=$HOME/autoware_map/sample-map-planning vehicle_model:=sample_vehicle sensor_model:=sample_sensor_kit