Argo v9-stable rev4

Argo Autonomous Sailboat

An autonomous sailboat system based on Dragonforce 65 hull, running on Orange Pi Zero 2W with ROS2. The system includes comprehensive sensor integration, autonomous navigation, and safety monitoring.

argo freecad 3d.webm

News

• May 2026: After ignominious series of problems at CCNW with cabling, water leaks, GPS and many other issues and only 30m of actual seawater sailing, came back to INI to fix these problems. Now IMU is working, GPS is working, dashboard has map display, there are new cable glands and a new hull that has been air-pressure tested for leaks. We are ready now for water trials again.
• Mar 2026: Version 4 of argo PCB with bug fixes, wind sensor on separate I2C bus, and new buzzer. Version 3 of the wind sensor PCB with RGBW LED controller. See Argo2: Robot Sailboat slides presented online at Embodied Intelligence Conference 20 March 2026.
• Feb 2026: Updated wind sensor design PCB and housing complete. 3D parts for wind sensor I2C cable and mast step up completed and printed.
• Jan 2026: Argo main PCB version 3 completed and seems fully functional with no surgery needed.
• Aug-Dec 2025: Intense development of new Argo PCB and ROS2 software framework. System services developed (see power_control/argo_power_control.py and launch/argo_lifecycle_manager.py) along with robust lifecycle management and battery/power control.
• Oct-Nov 2025: Developed first version of Argo simulator.
  • See simulator/sailboat-playground/README.md for usage and architecture.
  • Additional docs: Simulation Guide, Debug Simulation Guide

Installation

Pick Argo (boat) or Host (PC) below. ROS 2 must match your Ubuntu LTS: 22.04 (Jammy) → Humble; 24.04 (Noble) → Jazzy. Use the official ROS 2 Ubuntu (deb) install page that corresponds to your ROS release.

Argo (boat — Orange Pi Zero 2W)

On the robot, the stack targets Ubuntu 22.04–based Armbian and ROS 2 Humble. Bring up I²C/UART/PWM overlays, the radio/servo kernel module, Python dependencies, systemd services, and power control as described in the launch tree.

cd ~/argo   # or your clone path
sudo apt update && sudo apt upgrade -y
# Install ROS 2 Humble (.deb), then in each shell you use for builds:
#   source /opt/ros/humble/setup.bash

make install-all          # Python deps + optional hardware (see Makefile / messages)
make install-argo-cli
source ~/.bashrc
make -C launch install    # install systemd units; enable per docs

First-time SD card and hardware bring-up

1. Flash Orange Pi OS using Orange Pi Imager or equivalent.

2. Initial system setup

sudo apt update && sudo apt upgrade -y
sudo apt install python3-pip python3-dev build-essential
sudo apt install i2c-tools device-tree-compiler git
sudo apt install ros-humble-desktop

3. Clone repository

cd /home/orangepi
git clone https://github.com/SensorsINI/argo.git
cd argo

4. Python and ROS helpers

make install-python-deps   # requirements.txt on the boat
# If needed:
sudo apt install python3-rclpy python3-std-msgs python3-geometry-msgs

5. Hardware configuration

Edit /boot/orangepiEnv.txt and add overlays:

overlays=pi-i2c0 disable-uart0 ph-uart5 pi-pwm2 pi-pwm4
user_overlays=argo_radio_servo_overlay

Build and load the PWM capture module:

cd ~/argo/nodes/pwm_capture_module
make all
sudo modprobe argo_radio_servo_module   # or reboot

Add your user to hardware groups:

sudo usermod -a -G i2c,dialout $USER
# log out and back in

6. Verify

sudo i2cdetect -y 0
lsmod | grep argo
ls -la /sys/kernel/argo_radio_servo/
sudo cat /dev/ttyS5

7. CLI and services

make install-argo-cli && source ~/.bashrc
make -C launch install
sudo systemctl daemon-reload
sudo systemctl enable argo_launch.service

Host (simulation, Foxglove, remote monitoring)

Use this on a laptop or workstation for local simulation (asim), Foxglove Studio (same Wi‑Fi, VPN, or SSH tunnel), log playback, and shore-side scripts. The boat still runs ROS 2 on the Orange Pi; your host subscribes/records like any ROS 2 client. LoRa is a boat link—the monitoring machine uses normal network ROS 2 or recorded bags unless you run a separate LoRa gateway stack.

Prerequisites: Ubuntu 22.04 or 24.04 (see mapping to Humble/Jazzy above), git, make, sudo, and network access for apt.

One-shot bootstrap (Ubuntu 22.04 or 24.04):

cd /path/to/argo
make host-setup
source ~/.bashrc                     # reload Argo dotfiles in this terminal (or open a new tab)
# Dotfiles auto-source the first ROS found under /opt/ros (jazzy, iron, humble, rolling) when ROS_DISTRO is unset.

source .venv/bin/activate            # optional; pip tooling (argcomplete, pyglet, …)
asim

If a command in this terminal cannot find rclpy or ros2 right after host-setup, the ROS underlay is not loaded yet: run source ~/.bashrc (as above), or explicitly source /opt/ros/humble/setup.bash / source /opt/ros/jazzy/setup.bash matching your install.

make host-setup refuses to run on Orange Pi (use make install-all on the boat). It:

• Adds /etc/apt/sources.list.d/ros2-latest.list only if no file already references packages.ros.org/ros2/ubuntu (avoids a second Signed-By if you installed ROS 2 from the official doc first)
• Installs ros-<distro>-ros-base when /opt/ros/<distro> is absent (Jammy → humble, Noble → jazzy, or override with ROS_DISTRO=)
• Runs make install-deps (single apt-get update, then Argo’s ROS .deb set, Foxglove bridge, MCAP storage)
• Runs make setup-venv (creates .venv with uv if available, otherwise python3 -m venv, and installs requirements-host.txt) then make install-python-deps (same host requirements file again into .venv—idempotent on a PC)
• Runs make install-argo-cli and make submodule-init (install-argo-cli appends source <repo>/dotfiles/bashrc to ~/.bashrc; that file auto-sources /opt/ros/<distro>/setup.bash when ROS_DISTRO is not already set)

Tab completion (optional): Global argcomplete is not installed automatically on every shell (keeps login fast). After host-setup, run once: activate-global-python-argcomplete3 --user or activate-global-python-argcomplete --user, or see dotfiles/README.md — Tab Completion Setup.

Force a distro explicitly, e.g. on non-LTS or mixed images:

make host-setup ROS_DISTRO=jazzy

If apt-get update reports conflicting Signed-By for packages.ros.org/ros2/ubuntu, you have two ROS 2 source files—keep one. For example, if you already used the official installer, remove Argo’s duplicate and refresh: sudo rm -f /etc/apt/sources.list.d/ros2-latest.list && sudo apt-get update.

install-ros2-apt-repo checks for an existing ROS 2 source with sudo grep because some sources.list.d entries are root-only; without that, Make could miss your working repo and add a second file.

Foxglove: with ROS sourced on the machine that runs the bridge, start ros2 run foxglove_bridge foxglove_bridge and connect from Foxglove Studio to ws://<that-host-ip>:8765.

Makefile targets and shell aliases (reference)

• make host-setup — host-only full bootstrap (see above)
• make install-deps — ROS packages for the current ROS_DISTRO (Foxglove, MCAP, …)
• make install-python-deps — robot uses requirements.txt; host uses requirements-host.txt (prefers .venv if present)
• make install-hardware — PWM module (boat)
• make install-all — boat-oriented dependency + hardware path
• make -C launch … — lifecycle systemd helpers

After make install-argo-cli, common aliases include al / aq / ars (services), as / asq (status), asim (local sim), ar / ac (recording), alog, ah. Run make help and see dotfiles/.

System Overview

The Argo system consists of multiple ROS2 nodes that work together to provide autonomous sailing capabilities:

• Sensor Nodes: GPS (u-blox NEO-M9N), IMU (Adafruit BNO085), Wind sensors (3x Sensirion SDP3x), Battery/Water monitoring
• Control Interface: PWM capture for radio control and servo output
• Autonomous Control: Navigation and sail trimming algorithms
• Safety Systems: Manual override, battery monitoring, water intrusion detection

Demo Video

See the Argo autonomous sailboat in action at the 2024 CCNW (before current waterproofing and PCB developements): Argo Sailboat Demo

Hardware Platform

The main PCB design for Argo is available in the pcb/ folder. The latest board version is rev4, which contains the current sailboat control and sensor interface design.

• Orange Pi Zero 2W (Allwinner H618 SoC)
• GPS: u-blox NEO-M9N via UART5 (/dev/ttyS5)
• IMU: Adafruit BNO085 9-DOF Orientation IMU via I2C0 (0x4a)
• Wind Sensor: 3x Sensirion SDP3x differential pressure sensors via I2C0 (0x21, 0x22, 0x23). This PCB includes an RGBW LED and its I2C controller PCA9632 at individual address 0x62 (0x70 is LED All Call).
  • Design files: See pcb/WindSensor/WindSensor-rev1/ directory for full hardware schematics, layout, and BOM.
• ADC: MAX11612 for battery/water sensing via I2C0 (0x34)
• Environment: SHT45 temperature/humidity via I2C0 (0x44)
• PWM I/O: Custom kernel module for radio control and servo interfaces with high impedance safety mode
• LORA: Long range radio based on RA-01 radio module from ai-thinker.com using SX1276 LORA chip

Hardware Replacement and MAC Address Cloning

The Argo system is designed to handle frequent SBC hardware replacements (due to damage) while maintaining network identity. The WiFi MAC address is frozen to a consistent value (c8:26:e2:6c:58:ba) stored in network/ARGO_MAC_ID.txt and committed to git. This ensures:

• Network Stability: Same MAC address across hardware changes maintains network access (uzh-iot, ZeroTier, tobi-wlan)
• SD Card Portability: Same SD card works seamlessly with different SBC hardware
• Automatic Configuration: Run make freeze-mac-address after hardware replacement to restore network identity

See Network Improvements Documentation for details.

Servo Control Hardware

The Argo system uses a shared control circuit design that allows both radio control and autonomous software control of servos through a fail-safe high impedance switching mechanism:

Stock DF65 V6 (2018) Servo Configuration

• Sail Winch Servo: Joysway #880545

  • Type: Standard sail winch servo with high-precision plastic gears
  • Operating voltage: 4.8V to 6.0V
  • PWM frequency: 1520μs/50Hz
  • Input impedance: 12kΩ (measured)
  • Circuit resistor: 3.3kΩ (optimized for proper signal level)

• Rudder Servo: Joysway #881504

  • Type: Digital metal gear servo
  • Operating voltage: 4.8V to 6.0V
  • PWM frequency: 1520μs/330Hz
  • Input impedance: 400kΩ (measured)
  • Circuit resistor: 10kΩ (standard value works well)

Shared Control Circuit Design

The hardware implements a fail-safe shared control circuit where:

1. High Impedance Safety Mode: When PWM outputs are disabled (high impedance), radio control signals pass through resistors directly to servo inputs
2. Software Control Mode: When PWM outputs are enabled, software generates servo control signals
3. Voltage Divider Considerations: Resistor values are optimized based on servo input impedance:
   • Sail servo: 3.3V × (12kΩ / (3.3kΩ + 12kΩ)) = 2.6V (adequate signal level)
   • Rudder servo: 3.3V × (400kΩ / (10kΩ + 400kΩ)) = 3.2V (excellent signal level)

GPIO Pin Assignments (Orange Pi Zero 2W)

• PI11 (Pin 7): Radio Rudder Input - GPIO input with interrupt capability
• PI13 (Pin 26): Radio Sail Input - GPIO input with interrupt capability
• PI12 (Pin 33): Servo Rudder Output - PWM2 channel, high impedance when disabled
• PI14 (Pin 16): Servo Sail Output - PWM4 channel, high impedance when disabled

This design ensures radio control always works when software is not running, providing maximum safety for autonomous sailboat operation.

ROS2 Software Architecture

flowchart TB
  subgraph OUTPUTS[Output Nodes]
    RSR[rudder_sail_radio primary actuation output]
    LORA[lora long range communications]
  end

  CTRL[controller primary decision and control logic]

  subgraph SAFETY[Critical Safety Services]
    BW[battery_water critical]
    LA[launch critical]
  end

  subgraph SENSORS[Sensor Input Layer]
    GPS[gps]
    IMU[imu BNO085]
    ANEM[anem]
  end

  VR[Visualization and Recording]

  GPS --> CTRL
  IMU --> CTRL
  ANEM --> CTRL

Project Organization

The Argo system follows a modular ROS2 architecture with clear separation of concerns:

Directory Structure:

• nodes/ - Individual ROS2 nodes (Python-based sensor interfaces)
• launch/ - Lifecycle management, systemd services, and launch configurations
• power_control/ - Power management system (separate ROS2 package)
• foxglove/ - Visualization layouts for Foxglove Studio
• simulator/ - Sailboat simulation submodule (sailboat-playground)

Other READMEs

Core System Documentation

• Systemd Services Architecture - Service dependencies, boot sequence, and troubleshooting
• Network Improvements - WiFi reconnection, MAC address cloning, and NetworkManager configuration
• ZeroTier VPN Setup - ZeroTier VPN configuration, SSH access, and network management
• Power Control System - Power button, LED control, and graceful shutdown
• Systemd Integration - Launch system and lifecycle management
• Shutdown Fixes - Shutdown procedure fixes and improvements
• LED Status Guide - LED status display and hatch cover indicators

Development & Simulation

• Simulation System Overview - Sailboat simulation framework
• Simulation Documentation - Simulation setup and usage
• Debug Simulation - Debugging simulation issues

Hardware & Configuration

• I2C Configuration - I2C bus setup and troubleshooting
• LoRa (onboard ↔ shore) - SX1276 / Waveshare hardware, link parameters, packet header
• IMU BNO085 - C++ driver, Python bridge, calibration, argo_bno085 service
• Thermal monitoring (SBC) - Orange Pi GPU/CPU/DDR temperature logging (argo_thermal_monitor)
• Watchdog Configuration - Watchdog setup for Orange Pi Zero 2W
• CPU Frequency Tuning - CPU governor configuration
• Battery & Power Monitoring - Battery monitoring system
• Battery Panel Setup - Battery panel configuration
• SD Card Backup - SD card backup procedures
• DVFS Userspace Access - Dynamic voltage and frequency scaling

Visualization & Analysis

• Foxglove Visualization - Real-time visualization setup
• Foxglove Debugging - Debugging Foxglove issues
• Foxglove Joystick Setup - Joystick control configuration
• 3D Visualization - 3D visualization configuration
• Visualization Marker Persistence - Marker persistence in visualization
• Adding Sailing Maps - Map integration guide
• Maps Documentation - Sailing area maps

Remote Access & Interface

• Remote Desktop VNC - VNC remote desktop setup
• Web Dashboard - Mobile-friendly web interface

CLI & User Interface

• CLI Completion - Command-line completion setup
• Completion Quickstart - Quick start guide for CLI completion
• MOTD Customization - Message of the day customization

Additional Resources

• ROS2 Nodes - Node documentation and configuration
• Scripts - Utility scripts documentation
• Dotfiles Setup - CLI aliases and shell configuration
• Persistent Logging - Logging configuration

Simulation Support

Argo includes comprehensive simulation support for development and testing:

Local Simulation:

• Runs simulator directly on Orange Pi
• Uses sailboat-playground or mock simulator
• Excludes conflicting hardware nodes (GPS, IMU, anemometer, radio control)
• Perfect for algorithm development and testing

Remote Simulation:

• Offloads CPU-intensive simulation to remote machine
• SSH tunnel for ROS2 communication
• Centralized configuration management
• Ideal for resource-constrained environments

Quick Start:

# Local simulation
python3 launch/argo_lifecycle_manager.py simulate_local

# Remote simulation (requires setup)
python3 launch/argo_lifecycle_manager.py simulate_remote

# Or use helper scripts (equivalent flow):
./scripts/remote_simulator_tunnel.sh &
python3 scripts/remote_simulator_launch.py &
./scripts/launch_simulator_remote.sh

Directory Overview

launch/ - System Management & Services

Central control hub for the entire Argo system:

• argo_lifecycle_manager.py - Core lifecycle management with intelligent monitoring
• argo_*.sh - Shell scripts for start/stop/restart/status operations
• argo-launch.service - Systemd service configuration
• argo_gui.py - GTK-based status monitoring GUI
• argo_storage_monitor.py - Storage space monitoring and notifications
• Makefile - Service installation and management automation

nodes/ - ROS2 Node Implementations

Hardware interface and control nodes:

• Sensor Nodes: gps.py, bno085.py, anem.py, argo_battery_water.py, temp_monitor.py
• Control Nodes: rudder_sail_radio.py, controller.py, record.py
• pwm_capture_module/ - Custom kernel module for radio control and servo interfaces
• Configuration files: argo.yaml, calibration data, and support utilities

power_control/ - Hardware Power Management

Standalone ROS2 package for power button and LED control:

• argo_power_control.py - Main power control node with GPIO management
• Hardware Functions: Power button monitoring, LED patterns (heartbeat/SOS), graceful shutdown
• ROS2 Services: LED control, system health monitoring
• Makefile - Independent installation and service management

pcb/ - Hardware Design Files

Custom PCB development for production-ready integration:

• argo-v9-stable/ - Current stable PCB design (KiCad project files)
• datasheets/ - Component specifications and reference materials
• orange-pi/ - Orange Pi Zero 2W integration documentation and pin definitions
• Bill of Materials: Component sourcing and assembly documentation

foxglove/ - Real-time Visualization

Foxglove Studio integration for live system monitoring:

• argo_ros2.json - Pre-configured dashboard layout for Argo sailboat
• custom-argo-panel/ - TypeScript custom panel for specialized boat visualization
• setup_foxglove.sh - Automated Foxglove Bridge configuration and startup

system-monitoring/ - Optional System Monitoring Services

Optional system-level monitoring services for debugging and development:

• Makefile - Installation and management of system monitoring services
• services/ - Systemd service files for system monitoring
• scripts/ - Monitoring scripts for boot history, memory, processes, etc.
• Not installed by default - Use make install-system-monitoring to install

Core ROS2 Nodes:

• gps.py - GPS receiver interface (UART5, u-blox NEO-M9N)
• bno085.py - 9-DOF Orientation IMU with sensor fusion (I2C, Adafruit BNO085)
• anem.py - Wind sensor array (3x SDP3x pressure sensors)
• argo_battery_water.py - Power monitoring and safety systems
• rudder_sail_radio.py - Radio control input and servo output interface
• controller.py - Autonomous navigation and sail control algorithms
• record.py - Data recording management (ROS2 bag files)
• temp_monitor.py - System temperature monitoring
• argo_unified_simulator_bridge.py - Unified simulator bridge for local and remote simulation

Node Lifecycle Management

The Argo system uses a sophisticated lifecycle management approach centered around argo_lifecycle_manager.py:

Key Features:

• Real-time Monitoring: Active process detection during startup stabilization (not static sleeps)
• Failure Detection: Immediate FATAL error detection and reporting from systemd journal
• Auto-restart: Configurable restart policies with exponential backoff
• Graceful Shutdown: Proper cleanup and process termination
• Status Reporting: Comprehensive system health monitoring

Lifecycle Management Modes:

python3 launch/argo_lifecycle_manager.py run             # Launch all nodes
python3 launch/argo_lifecycle_manager.py stop            # Graceful shutdown
python3 launch/argo_lifecycle_manager.py restart         # Restart all nodes
python3 launch/argo_lifecycle_manager.py status          # Show system status
python3 launch/argo_lifecycle_manager.py monitor         # Continuous monitoring
python3 launch/argo_lifecycle_manager.py simulate_local  # Local simulation mode
python3 launch/argo_lifecycle_manager.py simulate_remote # Remote simulation mode

Critical Node Management:

• Critical Nodes: rudder_sail_radio.py, controller.py (essential for boat operation)
• Success Criteria: All critical nodes + minimum 3 total nodes running
• Failure Handling: Intelligent restart with failure analysis and error reporting

Startup Monitoring Pattern:

1. Launch Phase - Start all node processes
2. Detection Phase - Wait for nodes to register (30s timeout)
3. Stabilization Phase - Active monitoring for failures (15s with 1s intervals)
4. Validation Phase - Final status check and success determination

Systemd Integration:

• Service: argo-launch.service runs lifecycle manager in continuous mode
• Dependencies: Waits for network and hardware module initialization
• Restart Policy: Automatic restart on failure with 5-second delay
• Environment: ROS2 Humble sourcing and logging configuration

Running the System

Manual Launch (Recommended for Testing)

cd ~/argo
source /opt/ros/humble/setup.bash
python3 launch/argo_lifecycle_manager.py run

Individual Node Testing

# Test individual sensors with debug output
python3 nodes/gps.py --debug
python3 nodes/bno085.py bridge  # IMU bridge mode
python3 nodes/bno085.py status  # Check IMU health
python3 nodes/anem.py --debug
python3 nodes/argo_battery_water.py --debug
python3 nodes/rudder_sail_radio.py
python3 nodes/controller.py

System Monitoring

# Monitor ROS2 topics
ros2 topic list
ros2 topic echo /battery_voltage
ros2 topic echo /anem_speed_angle_temp
ros2 topic echo /rudder_sail_radio

Key ROS2 Nodes

• gps.py: GPS interface via UART5, publishes /gps_data
• bno085.py: Adafruit BNO085 IMU with sensor fusion, publishes /compass, /pose, /accel, /gyro, /imu_health
• anem.py: Wind speed/direction from 3 pressure sensors, publishes /anem_speed_angle_temp
• argo_battery_water.py: Power and safety monitoring, publishes battery/water alerts
• rudder_sail_radio.py: Radio control interface and servo output
• controller.py: Autonomous navigation controller
• argo_unified_simulator_bridge.py: Unified simulator bridge for local and remote simulation
• lora.py: Long range radio node for LORA module

Configuration Files

• argo.yaml: Main control parameters (mode, gains, etc.)
• argo.env: ROS2 environment variables for systemd services
• RTIMULib.ini: IMU calibration and sensor fusion settings

Safety Features

• High Impedance Safety Mode: Servo outputs default to high impedance (PWM disabled), allowing radio control to pass through directly to servos via resistor network
• Manual Override: Human can take control via radio at any time, with immediate priority over autonomous control
• Shared Control Circuit: Fail-safe design where radio control always works when software is not running
• Battery Monitoring: Automatic low battery alerts (7.2V threshold)
• Water Intrusion Detection: Immediate alerts on water sensor activation
• Sensor Fault Detection: Automatic reconnection and error handling
• Timeout Protection: Safe defaults if communication is lost

Data Analysis

To plot recorded bag file data, see argo-plots-2025.py (or argo-plots-2023.py).

Troubleshooting

Common Issues

1. I2C Permission Errors: Ensure user is in i2c group
2. UART Permission Errors: Ensure user is in dialout group
3. PWM Module Not Loading: Check device tree overlay installation
4. Sensor Not Found: Verify I2C connections and addresses
5. GPS No Data: Check UART5 overlay and baud rate settings
6. Servo Low Signal Levels: Check servo input impedance and adjust resistor values accordingly:
   • High-power servos (e.g., winch servos) may have low input impedance (12kΩ) requiring smaller resistors (3.3kΩ)
   • Standard digital servos typically have high input impedance (400kΩ) working well with 10kΩ resistors

Debug Commands

# Check system health
ah # show aliases
adevcheck # check for argo devices present and communicating
alog # follow argo logs - see help options
sudo i2cdetect -y 0 # check i2c bus
ros2 topic list
ros2 node list # check nodes
systemctl status argo-launch.service

2025 Development Updates

Exciting developments are underway for 2025:

• Custom PCB Design: A dedicated PCB is being designed to integrate all sensors and control systems into a single, robust board
• Enhanced Seaworthiness: Significant improvements to waterproofing, structural integrity, and marine-grade components for extended autonomous operation
• Production Readiness: Moving from prototype to production-ready autonomous sailboat system

Documentation

For detailed technical documentation, see the Google Doc README

License

BSD License - See package.xml for details