Modular SLAM (Python)

A modular SLAM (Simultaneous Localization and Mapping) implementation in Python for ROS 2.

Overview

This package provides a flexible SLAM system with both 2D and 3D capabilities, featuring:

• Front-end processing for both 2D and 3D LiDAR data
• Back-end optimization using pose graph optimization
• Loop closure detection using Scan Context

Package Structure

• mod_slam/ - Main SLAM implementation
  • front_end_2d.py - 2D LiDAR front-end processing
  • front_end_3d.py - 3D LiDAR front-end processing
  • back_end.py - Back-end optimization
  • utils/ - Utility modules for ICP, map management, scan context, etc.
• launch/ - ROS 2 launch files
• config/ - Configuration files for different robot platforms
• slam_interfaces/ - Custom ROS 2 message and service definitions

Dependencies

• ROS 2 (tested with Humble/Iron)
• Python 3.8+
• NumPy
• Open3D
• SciPy
• SmallGICP
• GTSAM

Installation

1. Clone this repository into your ROS 2 workspace:

cd ~/dev_ws/src
git clone https://github.com/Delport121/Modular_SLAM_py.git mod_slam_python

2. Install dependencies:

cd ~/dev_ws
rosdep install --from-paths src --ignore-src -r -y

3. Build the workspace:

colcon build --packages-select mod_slam slam_interfaces

4. Source the workspace:

source install/setup.bash

Usage

2D SLAM

ros2 launch mod_slam slam_2d.launch.py

3D SLAM

ros2 launch mod_slam slam_3d.launch.py

Voyager Platform

ros2 launch mod_slam slam_voyager.launch.py

Configuration

Configuration files are located in the config/ directory. You can modify parameters for:

• ICP matching
• Loop closure detection
• Map resolution
• Sensor topics

License

Authors

P.J. Delport

Acknowledgments