abougouffa / ndtpso_slam

ROS package for NDT-PSO, a 2D Laser scan matching algorithm for SLAM

NDTPSO-SLAM

NDTPSO-SLAM is a ROS package which provides an enhanced implementation of the scan matching algorithm proposed in ¹ and ². It uses NDT (Normal Distribution Transform) to model the environment and PSO (Particle Swarm Optimisation) to perform the scan matching optimization and solve the pose problem.

This package has been tested with ROS Melodic on Ubuntu 18.04, however; it should work with newer ROS1 versions.

This project is devided to two parts, the libndtpso_slam which is a C++ library that implements the proposed scan matching method, and in the other part, the ROS node ndtpso_slam_node which provides the ROS interface to this library.

Abdelhak BOUGOUFFA <abdelhak [dot] bougouffa [at] universite-paris-saclay.fr>

If you used this work, please cite:

• Sara BOURAINE, Abdelhak BOUGOUFFA and Ouahiba AZOUAOUI, Particle Swarm Optimization for Solving a Scan-Matching Problem Based on the Normal Distributions Transform, 10.1007/s12065-020-00545-y, Evolutionary Intelligence, Jan 2021. Download PDF ⏬
• Sara BOURAINE, Abdelhak BOUGOUFFA and Ouahiba AZOUAOUI, NDT-PSO, a New NDT based SLAM Approach using Particle Swarm Optimization, 10.1109/ICARCV50220.2020.9305519, 16th International Conference on Control, Automation, Robotics and Vision (ICARCV 2020), Dec 2020. Download PDF ⏬

Build NDTPSO-SLAM

This package has been tested on ROS Melodic and ROS Noetic

Dependencies

1. Eigen3 (can be installed from package manager)
2. OpenCV3+ (OPTIONAL: used to export map image)

To build NDTPSO-SLAM, clone it to your ROS Catkin workspace.

cd path/to/catkin_ws/src
git clone https://github.com/abougouffa/ndtpso_slam.git
cd ..
catkin_make --pkg ndtpso_slam

Using the provided node

You can edit the provided launch files to fit your LiDAR topic name and run:

roslaunch ndtpso_slam scan.launch

Footnotes

1. Sara BOURAINE, Abdelhak BOUGOUFFA and Ouahiba AZOUAOUI, Particle Swarm Optimization for Solving a Scan-Matching Problem Based on the Normal Distributions Transform, 10.1007/s12065-020-00545-y, Evolutionary Intelligence, Jan 2021. Download PDF ↩

2. Sara BOURAINE, Abdelhak BOUGOUFFA and Ouahiba AZOUAOUI, NDT-PSO, a New NDT based SLAM Approach using Particle Swarm Optimization, 10.1109/ICARCV50220.2020.9305519, 16th International Conference on Control, Automation, Robotics and Vision (ICARCV 2020), Dec 2020. Download PDF ↩

ndtpso_slam's Issues

Roadmap

1. Add map topic, publish the map
2. Check the published tfs and the pose

why the Average publish rate is so low?

[ INFO] [1676357464.295007732]: Average publish rate: 10.14Hz, matching rate: 4.27Hz
[ INFO] [1676357464.296691848]: laser_data.size = 1079
[ INFO] [1676357464.418336842]: Average publish rate: 8.46Hz, matching rate: 8.22Hz
[ INFO] [1676357464.418738750]: laser_data.size = 1079
[ INFO] [1676357464.571968731]: Average publish rate: 8.43Hz, matching rate: 6.53Hz
[ INFO] [1676357464.572164176]: laser_data.size = 1079