How open-rmf/rmf_demos works
Based on our analysis of the file tree and the README content, the project is a demonstration repository for the Open Robotics Middleware Framework (Open-RMF). Its purpose is to showcase how heterogeneous robot fleets can be coordinated through demos in different worlds (Hotel, Office, Airport Terminal, Clinic, Campus, Manufacturing & Logistics) using RMF’s scheduling, task allocation, and conflict resolution capabilities. It covers multiple integration aspects, from simulation launch files (both for Gazebo and Ignition), task dispatching, fleet adapters, and even bridges for external communication (like MQTT for RobotManager).
Main components and how they relate
• Demo Launch Environments: Multiple demo “worlds” (Hotel, Office, Airport Terminal, Clinic, Campus, Manufacturing & Logistics) that each have their own ROS 2 launch files and configuration directories.
• Simulation and Visualization: Integration with simulation tools (Gazebo, Ignition) and visualization components (RViz configuration files, schedule visualizer).
• Task Dispatching and Management: The rmf_demos_tasks directory includes numerous scripts for dispatching various task types (loop, clean, delivery, patrol, etc.). In addition, the README describes a bidding/dispatcher mechanism.
• Fleet Adapters and Bridges: rmf_demos_fleet_adapter and rmf_demos_bridges (which includes MQTT bridge and socketio bridge) are responsible for interfacing RMF with external simulated (or real) robot fleets.
• Maps and Assets: The rmf_demos_maps and rmf_demos_assets directories contain the environmental data and 3D models for robots and building structures used during the demonstrations.
• Web Integration: Although the primary demos are ROS launch-based, the README covers instructions to start a backend API server and a dashboard (rmf-web) via Docker – representing a web-based control/monitoring layer.
• Security and Advanced Features: Some demos integrate ROS 2 DDS-Security (for example, secure office world) and specialized features like lift watchdog, emergency alarms, and docking sequences.
Architectural patterns/design principles
• Layered Architecture: Clearly separate simulation/launch, task/mission management, fleet communication, and integration with external web monitors.
• Modular Deployment: Each “world” or demo is mostly independent, with its own configuration and launch files. Similarly, fleet adapters and bridges are standalone modules.
Relevant technologies and frameworks
• ROS 2 as the primary middleware, integrating with Gazebo (and Ignition via rmf_demos_gz).
• Docker for deployment of the backend API server and dashboard (rmf-web).
• MQTT and Socket.IO (in rmf_demos_bridges) for real-time communication to external systems.
• CMake (CMakeLists.txt files in several directories) and package.xml files for ROS package management.
• YAML configuration files across demos, which are essential for specifying robot behavior, building configurations, and potentially even the mapping information.
• Clear labels for all major components: Simulation Launcher, Task Dispatcher, Fleet Adapters, Bridges, Maps/Assets, and Web API/Dashboard.
• Annotations that show dependencies such as which modules require ROS 2, Docker, or Gazebo. Also include configuration layers (like YAML configuration files) that influence robot behavior in each demo.