This repository is the implementation of the open-sourced package, the GraphGNSSLib_LEO, which makes use of the LEO and GNSS coupled measurements to do positioning.
In this package, measured pseudorange and doppler shift of GNSS and simulated pseudorange and doppler shift of LEO satellites are used for positioning. Measurements from the historical and current epochs are structured into a factor graph which is then solved by non-linear optimization. Positioning results from single point positioning(SPP) and factor graph optimization (FGO) method are compared.
The package is based on GraphGNSSLib, adding modules to support simulated LEO satellites' measurements. This package is based on C++ which is compatible with the robot operation system (ROS) platform. Meanwhile, this package combines the RTKLIB (version: 2.4.3 b33) to read/decode the GNSS RINEX files. Users from Robotics field can easily have access to GNSS raw data for further study.
Important Notes:
- Be noted that the GNSS-only SPP mentioned throughout the package means estimating the positioing of the receiver based on the GNSS pseudorange Doppler measurements uisng SPP.
- Be noted that the GNSS-LEO SPPmentioned throughout the package means estimating the positioing of the receiver based on the GNSS and LEO pseudorange Doppler measurements uisng SPP.
- Be noted that the GNSS-LEO FGO mentioned throughout the package means estimating the positioing of the receiver based on the combination of GNSS and LEO pseudorange and Doppler measurements uisng FGO.
Authors: Yixin Gao,Weisong Wen, from the Trustworthy AI and Autonomous Systems (TAS) Laboratory, The Hong Kong Polytechnic University.
Related Papers:
-
Gao, Y., Wen, W., Liu, X., Huang F., Hu, J., & Su, D.(2025, Nov.) Enhancing LEO-GNSS Positioning in Urban Canyons: A Factor Graph Optimization Approach with StarLink Simulaieeetions. In 2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC). IEEE
-
Wen, W., & Hsu, L. T. (2021, May). Towards robust GNSS positioning and Real-time kinematic using factor graph optimization. In 2021 IEEE International Conference on Robotics and Automation (ICRA) (pp. 5884-5890). IEEE.
if you use GraphGNSSLib_LEO for your academic research, please cite our related papers
If you are not familiar with ROS, we highly recommend using our docker container to enjoy GraphGNSSLib_LEO. For the details, please go to the branch docker 4. Docker Support
Ubuntu 64-bit 18.04, ROS Melodic. ROS Installation. We only test it on Ubuntu 18.04 with ROS Melodic.
Follow the following instructions to install Ceres-solver instead of using the latest version of Ceres-solver.
Step 1: Download the Ceres-solver which is compatible with GraphGNSSLib_LEO.
Step 2: make and install
sudo apt-get install cmake
# google-glog + gflags
sudo apt-get install libgoogle-glog-dev
# BLAS & LAPACK
sudo apt-get install libatlas-base-dev
# Eigen3
sudo apt-get install libeigen3-dev
# make Ceres-solver
mkdir ceres-bin
cd ceres-bin
cmake ../ceres-solver
sudo make -j4
sudo make test
sudo make installsudo apt-get install ros-melodic-novatel-msgsClone the repository and catkin_make:
mkdir GraphGNSSLib_LEO/src
cd ~/GraphGNSSLib_LEO/src
mkdir result
git clone https://github.com/PolyU-TASLAB/GraphGNSSLib_LEO.git
cd ../
# if you fail in the last catkin_make, please source and catkin_make again
catkin_make
source ~/GraphGNSSLib_LEO/devel/setup.bash
catkin_make(if you fail in this step, try to find another computer with clean system or reinstall Ubuntu and ROS)
3.Run GNSS-LEO positioning using dataset collected in Whampoa UrbanNav-HK-Deep-Urban-1
The GNSS-LEO positioning via SPP and FGO is validated using static dataset collected near Whampoa of Hong Kong. Several parameters are as follows:
- GPS second span: 455342 to 456880
- satellite system: GPS/BeiDou/Stralink
- Window Size: Batch
- measurements considered: pseudorange and Doppler measurements
- result is saved by default
$(find global_fusion)/dataset/2021_0521_0607/GNSSLEO__SPP_trajectoryllh_psr_dop_fusion.csv
We provide some simulated LEO data, please enjoy it!
source ~/GraphGNSSLib_LEO/devel/setup.bash
# the GNSS-only SPP positioning result will be displayed in rviz
roslaunch global_fusion data_Whampoa_20210521_GNSS_only.launch
# You can record them as rosbag for usage in the following
rosbag record -aIf you want to use the LEO-GNSS integrated data, you can launch with the following steps:
source ~/GraphGNSSLib_LEO/devel/setup.bash
# run pseudorange and doppler fusion
roslaunch global_fusion data_Whampoa_20210521_GNSSLEO.launch
# You can also record all the topic for the following evaluation
rosbag record -a
# read GNSS raw data, LEO raw data and combine them as ROS topic
roslaunch global_fusion data_Whampoa_20210521_GNSSLEO.launch.launch
- GNSS only positioning using SPP with the blue arrow in topic /gnss_preprocessor_node/WLSENURTKLIB
The result is recorded in
GNSS_only_WLS_result.csv - GNSS only positioning using RTK with the red arrow in topic /gnss_preprocessor_node/ENUIntegerRTK
- GNSS and LEO positioning using SPP with the green arrow in topic /WLS_spp_psr
The result is recorded in
GNSSLEO__SPP_trajectoryllh_psr_dop_fusion.csv - GNSS and LEO positioning using FGO with purple curve in topic /FGOGlobalPath.
The result is recorded in
GNSSLEO_FGO_trajectoryllh_psr_dopp_fusion.csv
Please modify the file path for the result to suit your requirements.
Trajectories of three methods (GNSS-only SPP using pseudorange measurements with orange line, GNSS-LEO SPP positioning using pseudorange measurements with yellow line, and GNSS-LEO FGO positioning using pseudorange and doppler shift measurements with purple line. The x-axis and y-axis denote the east and north directions, respectively.
Positioning Error of Different method: Orange dots from GNSS-only SPP, Yellow dots from GNSS-LEO SPP, Purple dots from GNSS-LEO FGO.
To run GraphGNSSLib_LEO with docker, first make sure docker are installed on your machine. If you want to use the docker to run the global_fusion:
cd ~/catkin_ws/src/GraphGNSSLib_LEO/docker
make build
sudo -E ./start.bash #Do not delete " -E "
source devel/setup.bash
# run pseudorange and doppler fusion
roslaunch global_fusion data_Whampoa_20210521_GNSSLEO.launch
# you should open another ternimal to enter the docker.
# read GNSS raw data and publish as ROS topic
roslaunch global_fusion leo_raw_publisher.launchAlso, there is a video showing the demo after you have built the docker_file in the directory GraphGNSSLib/docker
If you want to restart the container, please stop it first:
sudo ./stop.bash
#then restart it
sudo -E ./start.bash The directory ~/shared_dir is created to connect the container and the host . In the container, it is located at ~/graph1/shared_dir, you can also download the code to shared_dir and compile the program in the container (Recommended for those who are interested in making changes to the source code)
We use Ceres-solver for non-linear optimization and RTKLIB for GNSS data decoding, etc. Some functions are originated from VINS-mono. The rviz_satellite is used for visualization. We based GNSS FGO proposed in GraphGNSSLib for positioning.
If there is any thing inappropriate, please contact Yixin GAO through yixin.gao@connect.polyu.hk or Weisong WEN through welson.wen@polyu.edu.hk.
The source code is released under GPLv3 license. We are still working on improving the code reliability.
For any technical issues, please contact Yixin GAO yixin.gao@connect.polyu.hk, from the Trustworthy AI and Autonomous Systems (TAS) Laboratory, The Hong Kong Polytechnic University.
For commercial inquiries, please contact Weisong WEN welson.wen@polyu.edu.hk, from the Trustworthy AI and Autonomous Systems (TAS) Laboratory, The Hong Kong Polytechnic University.