biomotion / baseline_localizer

This repo shows a simple way to utilize pcl:IterativeClosestPoint for localization

Baseline Localizer

Auther: Biomotion

Create Date: 2021/10/28

Last Modified: 2021/11/01

Dependencies

• geometry_msgs
• pcl_ros
• sensor_msgs
• tf2
• tf2_ros
• tf_conversions

Nodes

• pub_map

  • parameters: map_path
  • subscribe: N/A
  • publish: /map (sensor_msgs::PointCloud2)

• localizer

  • parameters: baselink2lidar_trans (float array), baselink2lidar_rot (float array), result_save_path (string)
  • subscribe: /map (sensor_msgs::PointCloud2), /lidar_points (sensor_msgs::PointCloud2), /gps (geometry_msgs::PointStamped)
  • publish: /lidar_pose (geometry_msgs::PoseStamped), /transformed_points (sensor_msgs::PointCloud2)
  • output: result poses as csv file saved in result_save_path

How to Use

• prepare your data
• start launch file
• play rosbag with slow speed
• check your result

Prepare Data

First, you should download your pcd map to your computer and modify map_path in pub_map_node.cpp for bag from ITRI.

For nuscenes datasets, you should utilize map_tile_loader package and setup your nuscenes map files.

Launch File

Start launch file by,

roslaunch baseline_localizer itri.launch
# or
roslaunch baseline_localizer nuscenes.launch

Play Rosbag

Play rosbag with the following command.

rosbag play --pause -r 0.1 <your_bag_file>

You can pause/resume the playing process by pressing white space on the rosbag play terminal.

It is recommanded to pause a while when the localizer node is initializing.

Check Result

The results will be saved in results/ folder by default. Check if there are correct number of lines and upload your result to our competition servers.