How open-rmf/free_fleet works
Project Type and Purpose
• The Free Fleet project is an implementation of the Open-RMF Fleet Adapter. Its main aim is to bridge communication between robot navigation stacks (using ROS 1 and ROS 2) and RMF (Robot Management Framework) core systems using zenoh as the messaging layer.
• The system supports different simulation setups (single/multi-robot) and uses docker containers for integration tests and simulations. The README explains multiple simulation examples (Nav1, Nav2) and required bridging between robot namespaces and the fleet adapter.
File Structure
• Top-level directories include:
• .github: Contains GitHub actions, workflows, and docker configurations for continuous integration and integration testing.
• free_fleet: Contains the core python implementation (conversion utilities, ROS type helpers, adapter utilities, and tests).
• free_fleet_adapter: Contains ROS adapter-specific modules (robot adapters for nav1, nav2, fleet adapter logic) along with launch files, tests, and a CMakeLists.txt. This layer bridges between the ROS navigation stacks and the Free Fleet core.
• free_fleet_examples: Provides sample configurations, launch simulations, maps, and example scripts to drive navigation commands – illustrating how the fleet adapter might be used in real-life scenarios.
• The presence of Dockerfiles and docker-compose files indicates that testing, integration, and different environments (ROS 1/ROS 2, minimal bridges, etc.) are part of the architectural design.
README for Additional Insights
• The README confirms the project uses zenoh bridges (one for ROS2 and one for ROS1) along with a dedicated zenoh router, emphasizing the separation of concerns:
• The robots run a non-namespaced navigation stack.
• Zenoh bridges, configured with each robot’s name as a namespace, route messages.
• The free_fleet_adapter then uses these messages to interface with RMF.
• Different simulation examples illustrate complexity like multiple robot fleets and varying integration scenarios (e.g., nav2 single vs. nav2 multiple robots, and nav1 setups).
• The README also provides detailed instructions for setting up dependencies and communicating with external services like rmf-web for task dispatching.
a. Main Components to Include:
• Robots/Navigation Stacks: Indicate both ROS1 (Nav1) and ROS2 (Nav2) stacked robots.
• Zenoh Bridges: Two bridges (zenoh-bridge-ros2dds for ROS2 and zenoh-bridge-ros1 for ROS1) that encapsulate the communication between robots and the adapter.
• Zenoh Router: Centralized message broker for routing zenoh messages.
• Free Fleet Adapter: The adapter logic (free_fleet_adapter) that receives messages from the various bridges and communicates with the RMF core.
• Free Fleet Core Library: The “free_fleet” python package, handling data conversion and utility functions.
• RMF Core / External Services: The RMF core system (or rmf-web API server) that dispatches tasks and controls fleet behavior.
• Simulation/Example Environment: Containers or simulation launch environments detailed under free_fleet_examples and the Dockerfiles.
b. Relationships and Data Flow:
• That the zenoh bridges publish their messages to the zenoh router.
• That the free_fleet_adapter fetches messages from the zenoh router.
• Interaction between RMF core (or rmf-web) and the free_fleet_adapter for task dispatching and fleet command exchanges.
c. Architectural Patterns and Design Principles:
• Decoupling between the robotics middleware (ROS 1/ROS 2 navigation stacks) and the fleet management system by using the zenoh communication layer.
• Modularity by separating the free_fleet core (utilities and converters) from the free_fleet_adapter (integration layer handling robot specific logic).
• Use of containerization (Docker) for integration tests and deployment setups, which is critical for ensuring consistent development and deployment environments.
d. Technologies, Frameworks, and Libraries:
• ROS 2 Jazzy and ROS 1 Noetic as supported middleware.
• Use of zenoh (including zenoh-bridge-ros2dds and zenoh-bridge-ros1) in the communication layer.
• Mention integration with rmf-core components and ancillary tools such as rmf-web, rmw-cyclonedds, and colcon for build orchestration.
• Dependencies like Python, CMake (for free_fleet_adapter), and Docker for simulations.