NCRL_LIO

Arthors

• Chun-Jung Lin - chadlin.gdr07g@nctu.edu.tw
• Jung-Cheng Yang - johnsongash.gdr07g@nctu.edu.tw
• Yin-Long Yan - yanlong658.gdr08g@nctu.edu.tw
• Bing-Yuan You - physical31031.c@nycu.edu.tw

Paper

A real-time LiDAR inertial odometer system (RTLIO) is developed to achieve high-precision and high-frequency odometry and mapping for feedback control of UAVs. In contrast to the traditional LIO approach, the initialization process in this work can be performed even when the device is stationary. The developed RTLIO can generate LiDAR-inertial odometery by solving cost functions that consist of the LiDAR and IMU residuals. Loop closure and pose-graph optimization are also developed in RTLIO to further reduce the accumulated pose errors. To demonstrate real-time feedback control of UAVs, RTLIO is applied to provide high frequency pose feedback for controlling an unmanned aerial vehicle (UAV). After long-range experiments, RTLIO can outperform LIO with a smaller drift. Experiments with publicly available dataset (i.e., KITTI1 ) are also conducted to demonstrate the efficacy of the algorithm in both accuracy and frequency.

Result

Indoor Experiment

KITTI Dataset

Prerequisites

1. ROS Kinetic or Melodic ROS INSTALLATION
2. Ceres Solver CERES INSTALLATION
3. PCL PCL INSTALLATION

Build NCRL_LIO

1. $ cd [workspace]/src
2. $ git clone https://chunjonglin@bitbucket.org/chunjonglin/ncrl_lio.git
3. $ catkin_make
4. $ source [workspace]/devel/setup.bash or $ source [workspace]/devel/setup.zsh

Run NCRL_LIO

operate with VLP-16

1. modify the parameters in ncrl_lio/config/[filename.yaml]
2. roslaunch ncrl_lio run.launch

operate with KITTI dataset

1. modify the parameters in ncrl_lio/config/[filename.yaml]
2. roslaunch ncrl_lio KITTI.launch