Skip to content

[ROS2][feat] Get loop closure indices

Tong Fu requested to merge ROS2/feat/GetLoopInidces into feat/ROS2

Goal

Before, loop closure indices are provided by a csv file. It's difficult to know loop closure indices before a slam map is built. This MR aims to give two ways to obtain loop indices for users:

  • In rviz, click a point on the trajectory which is considered as the loop frame of the current frame.
    • Play slam and stop the bag when a loop closure is observed.
    • Click the position where forms a loop with current position
    • the current pose and the closest pose to clicked point are added into LoopDetections
  • Save trajectory with index information so that one can get loop closure indices with saved trajectory.
    • Play slam and stop the bag when a loop closure is observed.
    • Save trajectory
    • Open trajectory with a visual tool and get the loop closure indices
    • Create a csv file of loop closure indices. More than one loop can be added in the file

Changes

  • Add ClickedPointCallback function
  • Save trajectory with index term
  • Add helper functions ParseSentence and ParseCSV
  • Overload ReadCSV to check csv fields
  • Refact ReadPose to check csv fields

Results

Data

  • data 1: car(hesai)
  • data 2: Acquisition_SIDO/outdoor

Build

colcon build --base-paths slam/ros2_wrapping --cmake-args -DCMAKE_BUILD_TYPE=Release -DENABLE_g2o=ON

Launch file

  • Use `slam_velodyne.py' but comment velodyne points conversion part.
  • change "velodyne" to "hesai" in tf_base_to_velo_node

Parameters

  • loop_closure.enable_offset: true
  • n_threads: 4
  • distance_threshold: 0.1
  • leaf_size.edges: 0.1
  • leaf_size.planes: 0.1
  • decaying_threshold: 900
  • min_frames_per_voxel: 0
  • voxel_grid_resolution: 1.
  • neighbors_radius: 0.05
  • edge_depth_gap_threshold: 2.
  • edge_nb_gap_points: 20

Run

For car

ros2 run lidar_conversions generic_conversion_node \
ros2 launch ros2 launch lidar_slam slam_velodyne.py \
ros2 bag play --clock path/to/bag --remap /hesai/pandar:=/generic_points

If the data is a pcap of hesai

ros2 run lidar_conversions generic_conversion_node \
ros2 launch ros2 launch lidar_slam slam_velodyne.py \
ros2 launch hesai_ros_driver start.py # some parameters need to be changed in config file, such as frame_id, 
Edited by Tong Fu

Merge request reports

Loading