How uuvsimulator/uuv_simulator works
Project type and purpose
• This is a modular ROS/Gazebo simulation framework (“uuv_simulator”) for Unmanned Underwater Vehicles. It provides:
• Gazebo physics and sensor plugins (C++)
• ROS message/service definitions
• Vehicle/world models (URDF/Xacro, SDF)
• ROS controller nodes (Python/C++)
• Teleop and assistant scripts
• Tutorial/launch infrastructure
High‐level architectural components
a. Simulation Core
– Gazebo Simulator (physics engine + SDF world loader)
b. Gazebo Plugins
– Hydrodynamics & inertial plugins (uuv_gazebo_plugins)
– Thruster & fin plugins
– World plugins (underwater current models)
– Sensor plugins (IMU, DVL, GPS, pressure, RPT, camera)
– ROS interface plugins (uuv_gazebo_ros_plugins & uuv_sensor_ros_plugins)
c. Models
– Vehicle descriptions (URDF/Xacro, meshes) in uuv_descriptions
– World definitions & assets (uuv_gazebo_worlds)
d. ROS Communication
– Topics (state feedback, sensor data, thruster commands)
– Services (set parameters, get model properties, change sensor state)
– Message/service definitions (uuv_control_msgs, uuv_sensor_plugins_msgs, uuv_world_ros_plugins_msgs)
e. Control Layer (ROS nodes)
– Thruster manager (uuv_thruster_manager)
– Control allocator (uuv_auv_control_allocator)
– PID, sliding mode, geometric controllers (uuv_control_cascaded_pids, uuv_trajectory_control)
– Utility nodes (uuv_control_utils)
f. Teleop & Assistants
– Teleoperation scripts (uuv_teleop)
– Automation helpers & launch scripts (uuv_assistants)
g. Tutorials & Demos
– Pre‐configured launch files & example worlds/scripts (uuv_tutorials)
Relationships and data flows
• Gazebo loads world (SDF) and vehicle (URDF) models.
• Gazebo plugins compute physics (forces, currents, hydrodynamics) and publish sensor data to ROS topics via ROS‐interface plugins.
• Controller ROS nodes subscribe to sensor/state topics (and /tf transforms), compute control actions, and publish desired wrench/thruster commands.
• Thruster manager node converts high‐level wrench commands into individual thruster commands and publishes them to the thruster plugin topics.
• Teleop nodes generate setpoints (e.g. joystick → cmd_vel) feeding into controllers.
• Services allow dynamic reconfiguration (e.g. set current velocity, enable/disable sensors).
Key architectural patterns and principles
• ROS modularity: each package is a self‐contained component providing nodes, plugins, messages, launch files.
• Publisher/subscriber for streaming data; services for parameter/configuration.
• Gazebo plugin pattern: separate C++ plugins for dynamics, sensors, world, each exposing ROS interfaces.
• Clear separation of concerns: simulation (Gazebo + plugins) vs. control (ROS nodes) vs. utilities/launch.
• Use of standard ROS build tools (Catkin/CMake) and ROS message/service definitions for interoperability.
Technologies, frameworks, and libraries
• Gazebo (physics, rendering, plugin API)
• ROS (nodes, topics, services, launch)
• C++ (plugins, high‐performance code)
• Python (ROS scripts, controllers, helpers)
• Xacro/URDF (vehicle descriptions)
• SDF (world definitions)
• CasADi (optional in control allocator)