Tramcar LMS1xx Perception

This is part of our tramcar perception system. It generate a fused grid map for obstacles with four SICK LMS151.

Check out a video that shows all objects with the distance of less than 15 meters:

Prerequisites

Set up workspace and catkin

Regardless of your system you will need to do the following steps:

cd <catkin_ws>            # navigate to the workspace
mkdir src                 # create src dir if you don't have it already
cd src
# Now you just need to clone the repo:
git clone --recursive https://github.com/zhearing/lms1xx_perception.git
git submodule update --init --recursive

System requirements

Warning: The driver of LMS1xx is Not Available under Ubuntu 18.04, you need to fix the bug by yourself.

You will need ROS, CMake, Eigen3, Armadillo, RViz plugins. The following sections contain an installation command for various Ubuntu systems (click folds to expand):

Ubuntu 14.04

Install these packages:

sudo apt install -y ros-indigo-desktop-full libeigen3-dev libopenblas-dev liblapack-dev libarpack* libarmadillo*

Ubuntu 18.04

Install these packages:

sudo apt install -y ros-melodic-desktop-full ros-melodic-rviz*

You can easily install these packages using sudo bash install.sh. Script is include in install.sh.

How to build?

This is a catkin package. So we assume that the code is in a catkin workspace and CMake knows about the existence of Catkin. It should be already taken care of if you followed the instructions here. Then you can build it from the project folder:

catkin_make -DCMAKE_BUILD_TYPE=Release -j8

How to run?

Run online

roslaunch lms1xx_perception four_scanners.launch
roslaunch lms1xx_perception lms1xx_perception.launch

Run offline

Note: Don't forget to use simulation time when republishing from a bag file: Using Sim Time.

roslaunch lms1xx_perception lms1xx_perception.launch

Authors

• Zeyu Zhong - Initial work - The single SICK LMS1xx ROS driver is originally from @clearpathrobotics. The obstacle_detector package is originally from @tysik.

License

This project is licensed under the LGPL-3.0 License.