| ROS focal/noetic | ROS focal/foxy | Gitlab CI | Coverage |
|---|---|---|---|
 |
 |
Lanelet2 is a C++ library for handling map data in the context of automated driving. It is designed to utilize high-definition map data in order to efficiently handle the challenges posed to a vehicle in complex traffic scenarios. Flexibility and extensibility are some of the core principles to handle the upcoming challenges of future maps.
Features:
- 2D and 3D support
- Consistent modification: if one point is modified, all owning objects see the change
- Supports lane changes, routing through areas, etc.
- Separated routing for pedestrians, vehicles, bikes, etc.
- Many customization points to add new traffic rules, routing costs, parsers, etc.
- Simple convenience functions for common tasks when handling maps
- Accurate Projection between the lat/lon geographic world and local metric coordinates
- IO Interface for reading and writing e.g. osm data formats (this does not mean it can deal with osm maps)
- Python bindings for the whole C++ interface
- Boost Geometry support for all thinkable kinds of geometry calculations on map primitives
- Released under the BSD 3-Clause license
- Support for ROS1, ROS2, Docker and Conan (see instructions below)
Lanelet2 is the successor of the old liblanelet that was developed in 2013. If you know Lanelet1, you might be interested in reading this.
You can find more documentation in the individual packages and in doxygen comments. Here is an overview on the most important topics:
- Here is more information on the basic primitives that make up a Lanelet2 map.
- Read here for a primer on the software architecture of lanelet2.
- There is also some documentation on the geometry calculations you can do with lanelet2 primitives.
- If you are interested in Lanelet2's projections, you will find more here.
- To get more information on how to create valid maps, see here.
Lanelet2 has been released for ROS. Just install ros-[distribution]-lanelet2, e.g.:
sudo apt install ros-noetic-lanelet2
There is a Docker container from which you can test things out:
docker build -t lanelet2 . # builds a docker image named "lanelet2"
docker run -it --rm lanelet2:latest /bin/bash # starts the docker image
python -c "import lanelet2" # quick check to see everything is fine
The docker image contains a link to your local lanelet2, so you can work and see changes (almost) at the same time. Work with two screens, one local and one on docker. Make your code changes locally, then run again catkin build on docker to recompile the code (update python modules).
In case you want to build it in your own way (without the above Docker image) use these instructions.
Lanelet2 relies mainly on Catkin for building and is targeted towards Linux.
At least C++14 is required.
Besides Catkin, the dependencies are
Boost(from 1.58)eigen3mrt_cmake_modules, a CMake helper librarypugixml(for lanelet2_io)boost-python, python2 or python3(for lanelet2_python)geographiclib(for lanelet2_projection)rosbash(for lanelet2_examples)
For Ubuntu, the steps are the following:
- Set up ROS, and install at least
rospack,catkinandmrt_cmake_modules(e.g.ros-melodic-rospack,ros-melodic-catkin,ros-melodic-mrt-cmake-modules):
sudo apt-get install ros-melodic-rospack ros-melodic-catkin ros-melodic-mrt-cmake-modules
- Install the dependencies above:
sudo apt-get install libboost-dev libeigen3-dev libgeographic-dev libpugixml-dev libpython-dev libboost-python-dev python-catkin-toolsOn 16.04 and below, mrt_cmake_modules is not available in ROS and you have to clone it into your workspace (git clone https://github.com/KIT-MRT/mrt_cmake_modules.git).
As usual with Catkin, after you have sourced the ros installation, you have to create a workspace and clone all required packages there. Then you can build.
source /opt/ros/$ROS_DISTRO/setup.bash
mkdir catkin_ws && cd catkin_ws && mkdir src
catkin init
catkin config --cmake-args -DCMAKE_BUILD_TYPE=RelWithDebInfo # build in release mode (or whatever you prefer)
cd src
git clone https://github.com/fzi-forschungszentrum-informatik/lanelet2.git
cd ..
catkin buildIf unsure, see the Dockerfile or the travis build log. It shows the full installation process, with subsequent build and test based on a docker image with a clean Ubuntu installation.
For non-catkin users, we also offer a conan based install process. Its experimental and might not work on all platforms, especially Windows. Since conan handles installing all C++ dependencies, all you need is a cloned repository, conan itself and a few python dependencies:
pip install conan catkin_pkg numpy
conan remote add bincrafters https://bincrafters.jfrog.io/artifactory/api/conan/public-conan # required for python bindings
conan config set general.revisions_enabled=1 # requried to use bincrafters remote
git clone https://github.com/fzi-forschungszentrum-informatik/lanelet2.git
cd lanelet2From here, just use the default conan build/install procedure, e.g.:
conan source .
conan create . lanelet2/stable --build=missingDifferent from the conan defaults, we build lanelet2 and boost as shared libraries, because otherwise the lanelet2's plugin mechanisms as well as boost::python will fail. E.g. loading maps will not be possible and the python API will not be usable.
To be able to use the python bindings, you have to make conan export the PYTHONPATH for lanelet2:
conan install lanelet2/0.0.0@lanelet2/stable --build=missing -g virtualenv # replace 0.0.0 with the version shown by conan
source activate.sh
python -c "import lanelet2" # or whatever you want to do
source deactivate.shThe python bindings are build for your default python installation by default (which currently is python2 on most systems). To build for python3 instead of python2, create a python3 virtualenv before initializing the workspace with catkin init. The command python should point to python3.
After catkin init run catkin config --cmake-args -DCMAKE_BUILD_TYPE=RelWithDebInfo -DPYTHON_VERSION=3.6 to make sure that the correct python version is used. Then build and use as usual.
Note: With bionic and beyond, the apt package python3-catkin-tools conflicts with ROS melodic and should not be used. Use either the python2 version or use pip to install the python3 version.
Examples and common use cases in both C++ and Python can be found here.
- lanelet2 is the meta-package for the whole lanelet2 framework
- lanelet2_core implements the basic library with all the primitives, geometry calculations and the LanletMap object
- lanelet2_io is responsible for reading and writing lanelet maps
- lanelet2_traffic_rules provides support to interpret the traffic rules encoded in a map
- lanelet2_projection for projecting maps from WGS84 (lat/lon) to local metric coordinates
- lanelet2_routing implements the routing graph for routing or reachable set or queries as well as collision checking
- lanelet2_maps provides example maps and functionality to visualize and modify them easily in JOSM
- lanelet2_matching provides functions to determine in which lanelet an object is/could be currently located
- lanelet2_python implements the python interface for lanelet2
- lanelet2_validation provides checks to ensure a valid lanelet2 map
- lanelet2_examples contains tutorials for working with Lanelet2 in C++ and Python
If you are using Lanelet2 for scientific research, we would be pleased if you would cite our publication:
@inproceedings{poggenhans2018lanelet2,
title = {Lanelet2: A High-Definition Map Framework for the Future of Automated Driving},
author = {Poggenhans, Fabian and Pauls, Jan-Hendrik and Janosovits, Johannes and Orf, Stefan and Naumann, Maximilian and Kuhnt, Florian and Mayr, Matthias},
booktitle = {Proc.\ IEEE Intell.\ Trans.\ Syst.\ Conf.},
year = {2018},
address = {Hawaii, USA},
owner = {poggenhans},
month = {November},
Url={http://www.mrt.kit.edu/z/publ/download/2018/Poggenhans2018Lanelet2.pdf}
}