Skip to content

Introduction

RDSim

RDSim is a high-fidelity Robo Delivery Simulator designed for autonomous delivery systems.

  • Realism: RDSim integrates advanced SLAM, localization, planning, and control technologies within the Gazebo simulation environment.
  • Modularity: The simulator supports various robotic platforms and is designed to be easily extensible for different applications.
  • Autonomy: Features include MPPI-based motion planning, topology-based navigation, and GPS-IMU fusion localization for real-world deployment.
  • Multi-Sensor Support: RDSim enables sensor fusion with:
  • 3D LiDAR SLAM (e.g., Velodyne, Ouster)
  • IMU integration (e.g., Xsens, MicroStrain)
  • GPS localization (e.g., RTK GPS)
  • Ease of Use: RDSim offers Docker-based deployment, enabling a seamless setup for simulation and development.

Tested on Ubuntu 22.04 / 24.04 with ROS 2 Humble and Gazebo Fortress / Ignition.

Build Status GitHub


Demo

RDSim Overview


SLAM Integration using GLIM and HDL Localization


Topology-Based Navigation with MPPI Control


Outdoor Driving Simulation with GPS-IMU Fusion


Features and Updates

(2024.11) Ongoing Enhancements

  • ROS 2 control integration
  • Segmentation of drivable areas
  • Advanced behavior tree (BT) applications for decision-making

(2024.10) Advanced Planning and Control

  • MPPI controller tuning
  • Topology map navigation with Nav2
  • GPS + IMU + HDL-based relocalization
  • Topology Map Editor for flexible route planning

(2024.09) SLAM & Localization Improvements

  • 3D LiDAR SLAM (GLIM)
  • Nav2 package configuration
  • GLIM factor graph to PCD conversion
  • Localization with (HDL localization) & Nav2 integration

(2024.04) Initial Development

  • Based on CitySim
  • Basic robot control and navigation
  • Docker container support
  • Gazebo dynamic actor simulation (gazebo_sfm_plugin)

See more details in Extension Modules and Demo pages.