Lilo

LiLo: Lite 3D-Lidar Odometry

This code obtains the odometry of a UGV (BLUE) with the estimation of the position transform between two point-clouds (a current one with a previous one) obtained by a Velodyne VLP16 3D-lidar sensor.

This code is modified from FLOAM, which was based on LOAM and A-LOAM .

The contribution of this version of the code is that the point cloud map is not compiled and the location is generated using a configurable batch of data in the launch parameters. This means that odometry estimation times do not accumulate and are low.

Modifier: Edison P. Velasco Sánchez, Universidad de Alicante, Spain.

Requisites

• ROS Kinetic or Melodic
• Velodyne repository

  sudo apt-get install ros-melodic-velodyne-pointcloud
  

• PCL (Point Cloud Library)
• Hector trajectory server

  sudo apt-get install ros-melodic-hector-trajectory-server
  

• Ceres Solver

Ceres instalation

• Clone the repository

git clone https://ceres-solver.googlesource.com/ceres-solver

• Install all the dependencies of Ceres

sudo apt-get install cmake libgoogle-glog-dev libgflags-dev libatlas-base-dev libeigen3-dev libsuitesparse-dev -y

• Build, test, and install Ceres. (latest stable release)

tar zxf ceres-solver-2.0.0.tar.gz
mkdir ceres-bin
cd ceres-bin
cmake ../ceres-solver-2.0.0
make -j3
sudo make install

Clone repository

    cd ~/catkin_ws/src
    git clone https://github.com/EPVelasco/lilo.git
    cd ..
    catkin_make --only-pkg-with-deps lilo

Ros Launch

    roslaunch velodyne_pointcloud VLP16_points.launch
    roslaunch lilo lilo_velodyne.launch

Test