FAST-LIO SLAM Plugin

FAST-LIO (Fast LiDAR-Inertial Odometry) provides real-time state estimation and mapping using tightly-coupled LiDAR and IMU fusion.

Overview

FAST-LIO2 is a robust, low-drift odometry algorithm that:

Provides accurate 6-DOF pose estimation
Builds incrementally updated point cloud maps
Supports Livox and Ouster LiDAR types
Operates in real-time at high frequency

Supported LiDARs

Pre-configured for:

Livox Mid-360
Ouster OS-1-32

Configuration

Configuration Files

Location: raisin_fast_lio_plugin/config/

Available configurations:

mid360.yaml - Livox Mid-360
ouster32.yaml - Ouster OS-1-32

Example Configuration (Ouster)

common:
  lid_topic: /ouster/points
  imu_topic: /ouster/imu
  time_sync_en: false

preprocess:
  lidar_type: 3  # 1=Livox, 2=Velodyne, 3=Ouster
  scan_line: 32
  blind: 0.5

mapping:
  acc_cov: 0.1
  gyr_cov: 0.1
  extrinsic_est_en: false
  extrinsic_T: [0.002441, 0.009725, 0.030662]
  extrinsic_R: [-1.0, 0.0, 0.0, 0.0, -1.0, 0.0, 0.0, 0.0, 1.0]

publish:
  scan_publish_en: true
  dense_publish_en: true

pcd_save:
  pcd_save_en: true

Parameter Reference

Input Topics:

Parameter	Default	Description
lid_topic	/livox/lidar	Input point cloud topic
imu_topic	/livox/imu	Input IMU data topic
time_sync_en	false	Enable software time sync
time_offset_lidar_to_imu	0.0	Time offset between LiDAR and IMU (seconds)

LiDAR Preprocessing:

Parameter	Default	Description
lidar_type	1	LiDAR type: 1=Livox, 3=Ouster
scan_line	6	Number of scan lines
blind	4.0	Minimum range filter (meters)
point_filter_num	3	Downsample factor for points

State Estimation:

Parameter	Default	Description
acc_cov	0.1	Accelerometer noise covariance
gyr_cov	0.1	Gyroscope noise covariance
b_acc_cov	0.0001	Accelerometer bias drift covariance
b_gyr_cov	0.0001	Gyroscope bias drift covariance
fov_degree	360.0	Field of view (degrees)
det_range	150.0	Maximum detection range (meters)

LiDAR-to-IMU Transform:

Parameter	Type	Description
extrinsic_est_en	bool	Enable online extrinsic calibration
extrinsic_T	vector<double>	Translation [x, y, z] from LiDAR to IMU (meters)
extrinsic_R	vector<double>	Rotation matrix (row-major 3x3) from LiDAR to IMU

Output Configuration:

Parameter	Default	Description
path_en	false	Publish trajectory path
scan_publish_en	true	Publish aligned point clouds
dense_publish_en	true	Publish full dense map
scan_bodyframe_pub_en	true	Output scans in body frame

PCD Export:

Parameter	Default	Description
pcd_save_en	true	Enable map saving to PCD file
interval	-1	Frames per PCD file (-1 = all in one)
map_file_path	./test.pcd	Output file path

Message Topics

Publishers

Topic	Type	Description
/Odometry	nav_msgs/Odometry	Odometry in LiDAR frame
/Odometry/base	nav_msgs/Odometry	Odometry in IMU/base frame
/cloud_registered	sensor_msgs/PointCloud2	Aligned map points
/cloud_registered_body	sensor_msgs/PointCloud2	Points in body frame
/Laser_map	sensor_msgs/PointCloud2	Global accumulated map

Subscribers

Topic	Type	Description
<lid_topic>	sensor_msgs/PointCloud2	Input LiDAR point cloud
<imu_topic>	sensor_msgs/Imu	Input IMU data

Usage

Selecting Configuration

Choose the config file matching your LiDAR sensor (mid360.yaml or ouster32.yaml)
Verify lid_topic and imu_topic match your sensor driver output
Set lidar_type correctly (1=Livox, 3=Ouster)

Extrinsic Calibration

For best performance, calibrate LiDAR-IMU extrinsics:

Measure physical offset from LiDAR to IMU
Set extrinsic_T with translation in meters
Set extrinsic_R with rotation matrix
Or enable extrinsic_est_en for online calibration

Warning

Incorrect extrinsics will cause drift and map distortion

Saving Maps

To save point cloud maps:

Set pcd_save_en: true
Set map_file_path to desired output location
Map is saved when plugin unloads