hdl-localization-ROS2

ROS2 wrapper for hdl-localization package with docker

Result

HDL-localization-result.mp4

1. Build docker image with Dockerfile

Before you start hdl-localization with docker, you should install docker and nvidia-docker in your PC.

You could also make docker image directly with provieded Dockerfile.

Move the terminal path to /docker and execute the following command.

cd docker

docker build -t hdl-localization-ros2:latest .

hdl-localization-ros2:latest is just example of this docker image, you can replace it with the image name you want.

After the image is created, you can execute docker images command to view the following results from the terminal.

output :

REPOSITORY                  TAG                    IMAGE ID          CREATED             SIZE
hdl-localization-ros2       latest                    812ae31625b3   48 minutes ago    3.36GB

2. Make HDL-Localization-ROS2 docker container

When you create a docker container, you need several options to use the GUI and share folders.

First, you should enter the command below in the local terminal to enable docker to communicate with Xserver on the host.

xhost +local:docker

After that, make your own container with the command below.

nvidia-docker run --privileged -it \
           -e NVIDIA_DRIVER_CAPABILITIES=all \
           -e NVIDIA_VISIBLE_DEVICES=all \
           --volume=${hdl_localization_repo_root}:/root/workspace/src \
           --volume=/tmp/.X11-unix:/tmp/.X11-unix:rw \
           --net=host \
           --ipc=host \
           --name=${docker container name} \
           --env="DISPLAY=$DISPLAY" \
           ${docker image} /bin/bash

⚠️ You should change {hdl_localization_repo_root}, {docker container name}, {docker image} to suit your environment.

For example,

nvidia-docker run --privileged -it \
           -e NVIDIA_DRIVER_CAPABILITIES=all \
           -e NVIDIA_VISIBLE_DEVICES=all \
           --volume=/home/taeyoung/Desktop/hdl-localization-ROS2:/root/workspace/src \
           --volume=/tmp/.X11-unix:/tmp/.X11-unix:rw \
           --net=host \
           --ipc=host \
           --name=hdl-localization-ros2 \
           --env="DISPLAY=$DISPLAY" \
           --env="QT_X11_NO_MITSHM=1" \
           hdl-localization-ros2:latest /bin/bash

If you have successfully created the docker container, the terminal output will be similar to the below.

output :

================HDL localization ROS2 Docker Env Ready================
root@taeyoung-cilab:/root/workspace#

3. Make custom HDL-Localization-ROS2 launch file

When we created our docker container, we utilized the --volume option, so the files we modified locally are directly available inside docker.

All you need is .pcd file, lidar topic and imu topic!

See hdl_localization_turtlebot.launch.py and run the example bagfile to get a better understanding.

points_topic = LaunchConfiguration('points_topic', default='/velodyne_points')         # velodyne topic name
odom_child_frame_id = LaunchConfiguration('odom_child_frame_id', default='base_link')  # velodyne_points frame ID
imu_topic = LaunchConfiguration('imu_topic', default='/imu')                           # optional, you should know noise parameter
globalmap_pcd = DeclareLaunchArgument('globalmap_pcd', default_value='/root/workspace/src/hdl_localization/data/turtlebot3.pcd', description='Path to the global map PCD file')

Build and Run it!

Within Docker, we need 3 containers.
To access the same container, use the command below

docker exec -it hdl-localization-ros2 /bin/bash

Within the connected container,

source /opt/ros/foxy/setup.bash

root@taeyoung-cilab:~/workspace#

All should stay on the above path.

1st container (For hdl localization package)

colcon build

source install/setup.bash

ros2 launch hdl_localization hdl_localization_turtlebot.launch.py

2nd container (For rviz)

rviz2 -d src/hdl_localization/rviz/hdl_localization_ros2.rviz

3rd container (For sample bagfile)

ros2 bag play src/hdl_localization/sample-bag/subset/

TODO

Support hdl_global_localization ROS2 launch file
Code refactoring
Support ROS2 humble
Suuport NDT CUDA

Acknowledgement

We based ours on the following packages.
This repository just provides a detailed guide to using docker to reduce dependencies.

Lee-JaeWon / hdl-localization-ROS2