ReefGuard

ReefGuard turns buoys into digital sentinels. Our sensor-equipped buoy detects vessels using underwater acoustics and wireless signals, alerting authorities when boats enter restricted habitats.

Comment

About the Project

Marine Protected Areas (MPAs) are designed to protect fragile ecosystems like coral reefs and marine habitats, but enforcement is extremely difficult. Oceans cover vast areas, and rangers or coast guards cannot monitor every mile of water at all times. As a result, ships often enter restricted zones undetected, sometimes accidentally, sometimes illegally. These violations can damage coral reefs, disturb marine wildlife, and cause environmental harm that may take decades to recover.

Our team was inspired by this problem after learning that industrial fishing still occurs in about 47% of coastal marine protected areas worldwide. We wondered:

What if the ocean could monitor itself?

This idea led us to create ReefGuard, a smart buoy system that acts as an automated sentinel for protected waters.

ReefGuard combines two sensing systems:

1. Acoustic Detection

Acoustic detection uses a TinyML model trained to recognize ship engine sounds and distinguish them from ocean noise or marine life.

2. Wireless Proximity Detection

Wireless proximity detection measures the signal strength of nearby electronics to estimate how close a vessel is to the buoy.

In our prototype, we used Wi-Fi signals as a proxy for nearby vessel electronics. By measuring Wi-Fi signal strength (RSSI) and calibrating signal values, we estimate the distance of nearby devices using interpolation.

Signal strength roughly follows an inverse power relationship with distance:

$$ RSSI \propto \frac{1}{d^n} $$

where

  • $d$ = distance from the buoy
  • $n$ = environmental path-loss exponent (typically $2 \leq n \leq 4$)

In a real deployment, the same principle would apply to broader radio-frequency emissions, such as:

  • Engine electromagnetic interference (EMI)
  • AIS (Automatic Identification System) broadcasts

When a vessel approaches a protected zone, ReefGuard automatically logs the detection, estimates its distance from the buoy, and sends an alert to a monitoring dashboard. The system can also raise a warning flag on the buoy to signal that a vessel has entered a restricted area.

How We Built It

Our prototype uses a GL.iNet GL-MT300N-V2 (Mango) router running OpenWrt to detect nearby devices using Wi-Fi signal strength. Custom scripts run on the router to monitor RSSI values and trigger alerts when signals exceed a defined threshold.

We also trained a TinyML sound classification model designed to run on an Arduino Nano 33 BLE Sense, which contains a built-in microphone and is capable of running lightweight machine learning models directly on the device.

A separate Arduino Uno serves as the actuator controller for physical alerts such as:

  • LEDs
  • Buzzers
  • A servo-driven warning flag

A PC-based dashboard displays detections in real time, logging events and estimating the distance of nearby vessels.

What We Learned

Through this project we learned:

  • How to design TinyML models for embedded devices
  • How to work with wireless signal strength (RSSI) to estimate distance
  • How sensor fusion can combine acoustic and radio signals for better detection
  • How to design a distributed environmental monitoring system

We also learned how difficult it can be to build reliable sensing systems in noisy ocean environments.

Challenges We Faced

One of our biggest challenges was simulating real-world vessel detection signals. In a real deployment, vessels would emit radio-frequency signals such as AIS broadcasts or engine electromagnetic interference.

However, since we did not have access to specialized RF equipment, we used Wi-Fi signals as a practical proxy to demonstrate the detection principle.

Another challenge was training a reliable audio classification model that could distinguish between ship engines and other environmental sounds.

Despite these challenges, our prototype successfully demonstrates the core concept: an autonomous buoy that detects approaching vessels and alerts authorities in real time.

Future Impact

In the future, ReefGuard could be deployed as a network of low-cost smart buoys, providing continuous monitoring for marine protected areas around the world.

Built With

  • arduino-nano-333-ble
  • arduino-uno
  • gl-mt300n-v2
  • microphone
  • servo-motor
Share this project:

Updates