LOAM based localization

[ataparlar]

Checklist

• I've read the contribution guidelines.
• I've searched other issues and no duplicate issues were found.
• I've agreed with the maintainers that I can plan this task.

Blocked by

• Point cloud feature extraction for LOAM based localization #6741

Description

LOAM based localization feature provides more robust and reliable localization for an ego vehicle in tunnels. This is a different localization feature than the current ones. This task is getting done under the Autoware Labs for Localization & Mapping Tasks.

This task has a mapping part. The link to the LOAM mapping task:

• Point cloud feature extraction for LOAM based localization #6741

For the development and testing of the feature, edge and surface feature pcd maps of YTU campus will be published. Here is the link to the data:

• https://drive.google.com/file/d/1caxN-uhCW9Y_ihpqUadlQT7KF-b4EQWR/view?usp=drive_link

Purpose

Establishing localization for ego vehicles in LOAM feature maps. Make it a localization package in Autoware.

Possible approaches

• Use the feature layers coming from the LOAM continuously to localize with the vehicle.
• Optimize the output pose continuously.

Definition of done

• Create brach and create package for the task
• LOAM Based Localization Task 1 - Implement point cloud undistortion #7449
• LOAM Based Localization Task 2 - Create depth image from the point cloud (project point cloud) #7450
• LOAM Based Localization Task 3 - Extract LOAM features from the projected cloud #7451
• LOAM Based Localization Task 4 - Implement feature matching #7452
• LOAM Based Localization Task 5 - Implement Localization #7453

[ataparlar]

An issue for creating the point clouds is opened. Here is the link:

• Point cloud feature extraction for LOAM based localization #6741

Additionally, this comment contains the details for using the data:

• Point cloud feature extraction for LOAM based localization #6741 (comment)

[ataparlar]

Hi everyone,

I have uploaded some data for testing and developing the package. LOAM feature maps can be found in this link:

• https://drive.google.com/file/d/1caxN-uhCW9Y_ihpqUadlQT7KF-b4EQWR/view?usp=sharing
• The testing rosbag for this environment will be published later.

After we have complete the below task it can be used for this package and LOAM mapping too.

• Collect Urban Dataset for Autoware and Autoware Labs autoware#4725

[ataparlar]

I have created this fork and branch for this task:

• https://github.com/ataparlar/autoware.universe/tree/feat/loam-feature-localization

Currently working on this.

[stale]

This pull request has been automatically marked as stale because it has not had recent activity.

[ataparlar]

Steps to Run the Code

Prepare the Data & Build

• Download the point clouds data:

mkdir -p ~/loam_based_localization_data/point_clouds
gdown --id 1OkSN53n-3BjAcU_gLzs5Il8IOFsWTnIa -O ~/loam_based_localization_data/point_clouds/
gdown --id 1YOSu9Bp32XgO643jNkZ7lGuSCYOcigwU -O ~/loam_based_localization_data/point_clouds/

• Download the rosbag data (rosbag size is nearly 60gb):

mkdir -p ~/loam_based_localization_data/test_rosbag
gdown --id 1uta5Xr_ftV4jERxPNVqooDvWerK0dn89 -O ~/loam_based_localization_data/test_rosbag/ 
gdown --id 10tw3sBZzVAiu9gWbB4mMclGuzY2-86In -O ~/loam_based_localization_data/test_rosbag/  

• Install the dependencies:

sudo apt install ros-humble-perception-pcl \
  	   ros-humble-pcl-msgs \
  	   ros-humble-vision-opencv \
  	   ros-humble-xacro
sudo add-apt-repository ppa:borglab/gtsam-release-4.1
sudo apt install libgtsam-dev libgtsam-unstable-dev

• Create a workspace and src directory inside of it. Clone the local-map branch of the repository.

mkdir -p ~/loam_localization_ws/src/
cd ~/loam_localization_ws/src/
git clone -b local-map https://github.com/ataparlar/LIO-SAM.git

• You need to change the point cloud file path at the line 11 and 12 in the config/params.yaml file in the package.
  • For instance it must be:

    cornerPointCloudPath:  "/home/ataparlar/loam_based_localization_data/point_clouds/tunnel_local_corner.pcd"
    surfacePointCloudPath: "/home/ataparlar/loam_based_localization_data/point_clouds/tunnel_local_surface.pcd"
  

cd ~/loam_localization_ws/
colcon build --symlink-install --cmake-args -DCMAKE_BUILD_TYPE=Release

Run

• Source the environment and launch.

source ~/loam_localization_ws/install/setup.bash
ros2 launch lio_sam run.launch.py

• Run the rosbag in another terminal:

source ~/loam_localization_ws/install/setup.bash
cd ~/loam_based_localization_data/
ros2 bag play test_rosbag/ --remap /tf_static:=/asd --start-offset 711

[xmfcx]

@ataparlar I'm closing this task since it won't be worked on anymore. Please feel free to reopen if you want to work on it again.