Proof of Live current data

Drift-Based Ocean Transparency Platform

Inspiration

Problem Statement

Illegal fishing costs the global economy billions annually while vast areas of the ocean remain sparsely monitored. Traditional patrol vessels burn enormous amounts of fuel, and satellite systems cannot always verify vessel identity at sea level. Illegal, unreported, and unregulated (IUU) fishing removes billions of dollars of seafood from global oceans every year, threatening marine ecosystems, coastal economies, and global food security. In many regions, up to 30% of total catch is estimated to be illegal.

The challenge is not awareness — it is detection.

The ocean covers 70% of the planet, yet enforcement coverage is sparse and reactive. Many illegal vessels disable AIS tracking, spoof identities, or operate in remote waters beyond consistent patrol visibility. Existing monitoring systems rely heavily on satellite passes, self-reported data, or manual inspection, creating gaps where illegal activity can persist undetected.

Additionally, most monitoring frameworks treat vessels as isolated signals rather than dynamic actors moving within ocean currents, fuel constraints, and biological hotspots. This lack of physics-aware modeling limits the ability to distinguish suspicious behavior from normal maritime operations.

Without scalable, autonomous, and behavior-aware monitoring systems, illegal fishing continues to exploit the ocean’s size, motion, and jurisdictional complexity.

We asked a different question:

What if the ocean itself moved the monitoring system?

Instead of powering across the Pacific, our vessel rides major ocean currents, using minimal propulsion only for steering corrections. The ocean becomes propulsion. The platform becomes persistent.

The Core Concept

We built a current-driven autonomous surface vessel that drifts along major circulation routes while monitoring maritime activity.

The goal is not confrontation.
The goal is verification.

Our system cross-checks physical vessel markings against official databases of legal fishers and their flag states.

Mesh Network Protocol

Our system doesn’t rely on cell service or constant internet access to work. Instead, we deploy a fleet of small floating devices that communicate directly with each other, forming a mesh network — like a team passing notes across the water. When one float detects a boat, it captures the image and shares it with nearby floats. The message hops from one device to the next until it reaches any float that has an internet connection. That single connected float then sends the information to authorities or cloud systems. This means the network keeps working even in remote ocean areas with no traditional coverage, because the floats cooperate and relay messages for each other.

Engineering Design

Our engineer figured out a new way we can hop currents like lanes like no other boat does being more sustainable than anything before

Our Algorithm

Blue Vector runs a receding-horizon optimizer — every 6 hours it wakes up, looks at where the boat is, checks the
real ocean current and wind at that position, then plays a fast mental game: it tries 19 different steering angles (from 45° left of the target bearing to 45° right, in 5° steps) and for each one simulates 5 days into the future.
For each of those 19 futures it adds up a cost — motor energy burned, time elapsed, distance still remaining, and a penalty for being in a current lane that's pushing the wrong way. It picks the angle with the lowest cost, commits to that heading for the next 6 hours, then forgets everything and repeats. The boat actually moves from three forces at once: the ocean current pushing it for free, the wind filling the sail for free (35% efficiency), and the small motor which isn't there to cross the ocean — it's just a rudder, 0.3 m/s, only powerful enough to choose which current lane to be in. Because the motor is so weak relative to the currents and wind, the optimizer naturally learns to hunt for favorable lanes: a current flowing toward Los Angeles is worth steering into even if it curves the route north, because it provides velocity the motor never could. No rule says "find the Kuroshio" — that behavior falls out of the cost function automatically. The result across 10,000 km is that the ocean and wind do roughly 77% of the work, and the motor just steers.

more info https://github.com/MateDort/Blue_Vector/tree/claude?tab=readme-ov-file

Hull & Survivability

Low-drag displacement hull optimized for 2–4 knot efficiency
Stability for long-duration exposure to Pacific storm conditions
Anti-biofouling treatment for multi-month deployment
Sealed electronics and autonomous fault recovery systems

Electronics Stack

Compute Core: Raspberry Pi 5

Runs onboard computer vision models
Processes vessel imagery
Parses AIS signals
Handles data filtering and compression

Sensors:

High-resolution marine camera
AIS receiver
GPS
Passive acoustic sensor

Identity Verification System

Our vessel performs multi-layer verification:

Camera captures vessel imagery
Raspberry Pi 5 detects:
- Vessel name
- Registration number
- Flag state
OCR (optical character recognition) extracts identifying text
Data is cross-referenced with legal fishing registries
Flag state and registration are matched against official databases

If:

AIS identity matches visual registration → compliant
AIS absent but visual ID present → flagged for review
Visual ID inconsistent with registry → anomaly detected

Only relevant image frames and metadata are transmitted to the data center to minimize bandwidth and power use.

Data Flow Architecture

Local edge processing on Raspberry Pi 5
Suspicious detections flagged
Compressed images and metadata transmitted to centralized server
Server cross-checks:
- National registries
- Regional fisheries management organization databases
- AIS logs
Compliance reports generated