How scottmayberry/mur works
Project Type and Purpose
• MUR is an open-source underwater robotics platform combining custom hardware (CAD and PCBs), embedded firmware (Teensy controllers), and a ROSbased software stack running on a compute module.
• It targets precise underwater navigation, sensing, modeling, and control.
File Structure
• Top-level domains:
• hardware/ (CAD models, BOM, PCBs, assembly videos, firmware)
• software/ (ROS packages: mur, mur_sensors, mur_model, mur_control, BlueBuzz integration)
• images/ (renders, photos)
• hardware/firmware → divides between:
• Teensy embedded code (EthernetManager, SensorManager, ThrusterManager)
• Compute module setup scripts (camera streaming, broadcast IP)
• software/ROS_packages → four core ROS packages plus support tools.
Extract Key Architectural Insights from README
• The mur ROS package is the orchestrator (roslaunch mur.launch).
• Subpackaging:
• mur_sensors: collects and publishes sensor data over ROS topics.
• mur_model: builds the robot’s physical/dynamic model, publishes transforms and state.
• mur_control: runs manual and PID-based thruster control, subscribes to sensor/model topics and sends commands.
• BlueBuzz and camera streaming nodes provide communication and video feed.
• Communication: ROS topics and services; UDP between compute module and Teensy; acoustic (BlueBuzz), Ethernet, WiFi links.
Main Components (Canvas Elements)
• Hardware Layer (physical):
• Structural Frame, Thrusters, IMUs, Pressure Sensors, Leak Detectors, Temperature/Humidity Sensors.
• Communication Modules: BlueBuzz acoustic modem, Ethernet port, WiFi interface.
• Compute Module (e.g., Raspberry Pi/NUC) running Linux + ROS.
• Embedded Microcontroller (Teensy) connected to thrusters and sensors.
• Firmware Layer:
• Teensy Firmware: SensorManager, ThrusterManager, EthernetManager.
• Compute Module Startup & Service Scripts: broadcast_ip_data.py, udp_camera_streamer.py.
• Software Layer (ROS):
• roslaunch orchestrator (mur.launch).
• mur_sensors nodes → publish /imu, /depth, /leak_status, etc.
• mur_model nodes → subscribe sensor topics, publish /tf, odometry, simulated model state.
• mur_control nodes → subscribe model & manual input topics, publish thruster setpoints.
• BlueBuzz nodes → ros_bluebuzz_udp, ros_bluebuzz_transmit.
• Camera streaming node.
• Configuration Store:
• YAML files: global_config.yaml, environment_info.yaml, control_info.yaml.
• External Interfaces:
• Operator Console (keyboard, GUI).
• Video Monitoring (UDP/MJPEG stream).
• Data Logging (rosbag, onboard file system).
Relationships and Data Flows
• Hardware Sensors → Teensy Firmware (I²C/SPI/ADC) → EthernetManager (UDP) → ROS topics.
• Compute Module:
– ROS Master ↔ all ROS nodes (mur_sensors, mur_model, mur_control, BlueBuzz, camera).
– roslaunch loads nodes and YAML params.
• Control loop:
– mur_sensors → mur_model → mur_control → Teensy (thruster commands) → Thrusters.
• Camera Hardware → camera stream script → ROS topic or external video client.
• BlueBuzz modem ↔ ros_bluebuzz_udp ↔ mur package via ROS topics.
Architectural Patterns & Principles
• Modular ROS nodes—each package encapsulates one concern (sensors, model, control, comms).
• Publish/Subscribe decoupling.
• Configuration‐driven (YAML) for environment and control parameters.
• Embedded real‐time firmware offloads critical I/O and timing from compute module.
• Single entrypoint orchestration (mur.launch) for ease of deployment.