WATER2030 - Water Autonomous Treatment and Environmental Restoration

Project Evolution

From COVID-19 Response to Award-Winning Innovation

2020

Project Conception

WATER2030 was conceived during the COVID-19 pandemic, addressing the challenges authorities faced in maintaining and cleaning marine environments when human resources were limited.

It was mostly driven by Diomidis' deep love for the sea and desire to preserve and restore the environment with the aid of science/engineering.

2024

ERC Grant Proposal

Submitted as an ERC Consolidator Grant 2024 proposal to the European Research Council. Despite not securing funding, the proposal laid the groundwork for future developments.

2024

Prototype Success

The vision inspired Ernesto Sigalas at Technical University of Crete to develop a working prototype, winning first place at POPRI EUSAIR Youth 2024:

🏆 Team: Ernesto Sigalas, Marios Konstantoudakis
🎓 Mentor: Diomidis Katzourakis
📰 Featured in: CNN Greece, Local Media

Vision

Looking Forward

WATER2030 remains a comprehensive vision for autonomous marine cleanup. While the prototype demonstrates technical feasibility, the full scope of the project - including fleet operations, advanced AI integration, and widespread deployment - awaits realization through future research and funding opportunities.

The Plastic Pollution Crisis

Addressing the critical need for autonomous water cleanup solutions

Despite EU policy goals to prevent waste, plastic packaging waste continues to increase. The Mediterranean Sea faces a particular crisis, with nearly 80% of microplastic pollution originating from land sources.

700+

Tons of plastic entering Mediterranean daily

100B

Surface floating plastic items in Balearic islands study

50%

EU target reduction in marine plastic litter

188kg

Packaging waste per capita (2021)

Trash-waste drift in harbors and stagnated oil spills in busy coastal areas

Key Challenges

Increasing plastic waste generation despite prevention policies
Mediterranean Sea comparable to subtropical gyre accumulation zones
Limited attempts to reduce existing stock of maritime plastic litter
Need for cost-effective, scalable autonomous harvesting solutions

WATER2030 Solution

Autonomous water waste harvesting platform with minimal human supervision

🚁

TFV Platform

Prototype TFV with expandable architecture for autonomous waste harvesting

🏗️

Infrastructure

Develop supporting infrastructure including docking stations, communication networks, and user interfaces

📋

Operations

Operational guidelines, training protocols, and comprehensive legal framework study

3D concept design of the Trash Feaster Vessel (TFV) with autonomous navigation capabilities

Collaborative trash collecting with multiple TFVs and support infrastructure

Design Principles

🛡️ Safety First

Failsafe design with Return-to-Base functionality
Fault tolerant systems
COLREGs compliance
Wildlife protection measures

💰 Cost Effective

Target: <20k€ manufacturing cost
Low maintenance requirements
Minimal human supervision
Scalable production design

🔧 High Reliability

15,000 km operational lifespan
Redundant critical systems
75%+ uptime target
Smart diagnostics

Technical Architecture

Advanced autonomy stack with AI-powered perception and planning

TFV System architecture with redundancy in energy and communication domains

Autonomy stack architecture showing perception, behavior, and planning subsystems

Platform Specifications

🚢 Physical Design

Unsinkable catamaran hull design
<40kg empty weight, <2m² footprint
360° rotatable jet thrusters
Retractable sensor tower
Anti-capsize design

🧠 Autonomy Features

Computer vision waste detection
YOLO/R-CNN object recognition
Real-time path planning
GNSS + sensor fusion localization
Generative AI behavior modeling

⚡ Performance Targets

>0.4 m/s harvesting speed
15kg/50L waste capacity
3+ hours operation time
5km cruise range
Solar + battery power

Autonomy Operation Space (AOS)

Maturity Level	Wave Height	Wind Speed	Visibility	Environment
M1 (Initial)	<0.1m	<11 km/h	>500m	Marinas, harbors
M2 (Advanced)	<0.2m	<19 km/h	>500m	Docks, lakes, bays
M3 (Vision)	<0.3m	<19 km/h	>200m	Coastlines, estuaries

Development Approach

Phase 1

Requirements & Architecture

Define system requirements, develop platform architecture, and establish safety protocols.

Phase 2

Prototype Development

Build and test initial TFV prototype with basic autonomy capabilities in controlled environments.

Phase 3

AI Integration

Implement advanced perception, behavior modeling, and planning algorithms using machine learning.

Phase 4

Field Testing

Extensive testing in real marine environments with iterative improvements and validation.

Research Team

Expertise in autonomous systems, AI, and environmental engineering

Diomidis Katzourakis

Principal Investigator & Project Lead

See about Diomidis Katzourakis for full bio and contact details.

Pierre Gourdain

Associate Researcher - Stakeholder Management & Compliance

10+ years building 200m$+ revenue organizations in technology and logistics. Former French Ministry of Environment. Leading stakeholder management and regulatory compliance efforts.

📧 pierre.gourdain [at] gmail [dot] com
🎓 IEP Paris Master's, INSEAD MBA
🌍 Multilingual: English, German, French, Greek

Ernesto Sigalas

Lead Developer - TFV Prototype

Lead developer of the award-winning Trash Feaster Vessel prototype at Technical University of Crete. Successfully demonstrated autonomous waste collection capabilities in controlled environments.

Collaboration Network

WATER2030 emphasizes collaborative development with governmental and non-governmental organizations, academic institutions, and existing water surface cleaning industries. The vision is global deployment of TFV systems wherever suitable Autonomy Operation Spaces can be established.

Environmental Impact

Scalable solution for global water cleanup initiatives

Feasibility Example: Mediterranean Coastline

Parameter	Target Value	Impact
Coastline Coverage	46,000 km × 100m width	4,600 km² cleaning area
Fleet Size	~650 TFVs	Complete coverage in 4 years
Total Mission Cost	<40 million €	Cost-effective large-scale deployment
Manufacturing Target	<20k € per TFV	Scalable production economics

Key Innovations

Generative AI Integration: Advanced behavior modeling and explainable decision-making
Open Source Platform: Collaborative development model for global adoption
Systems Engineering: Proven automotive industry methodologies applied to marine robotics
Cost Optimization: Designed for mass production and minimal operational costs
Environmental Focus: Wildlife-safe design with ecosystem preservation priority

🎯 Primary Targets

River mouths and estuaries
Ports and marina areas
Tourist coastal zones
Natura2000 protected sites

🔬 Research Applications

Microplastic monitoring
Water quality assessment
Marine ecosystem mapping
Pollution source tracking

🌍 Global Vision

Open-source technology platform
International collaboration
Scalable manufacturing
Policy framework development

State of the Art

Building upon existing solutions with advanced autonomy

Current Industry Solutions

🌊 The Ocean Cleanup

Interceptor solutions for rivers
Solar-powered conveyor systems
Barrier and tender combinations
Focus on large-scale operations

🤖 Ranmarine WasteShark

Predefined path following
SharkPod docking concept
Commercial deployment
Limited autonomy features

🦾 Academic Research

Computer vision for waste detection
LiDAR obstacle avoidance
ROS-based control systems
Limited real-world deployment

WATER2030 Advantages

Advanced AI Integration: Generative AI for behavior modeling and decision explanation
Automotive-Grade Engineering: Proven systems engineering from autonomous vehicle industry
Cost-Optimized Design: Target manufacturing cost enables large-scale deployment
Fault-Tolerant Architecture: Multiple redundancy levels for reliable autonomous operation
Open Source Approach: Collaborative development for global impact

Technical Implementation

Advanced perception, planning, and control systems

Generative AI aided TFV ego behavior system with context translation and supervisory monitoring

Perception System

🔍 Computer Vision

YOLO object detection
Semantic segmentation
Real-time waste classification
Marine environment adaptation

📡 Sensor Fusion

Multi-camera arrays
LiDAR for obstacle detection
Underwater sonar systems
GNSS + IMU localization

🎯 Target Detection

Surface plastic detection (4m range)
Semi-submerged waste (2m range)
90% detection confidence
Microplastic filtering

Planning & Control

The TFV uses advanced path planning algorithms combining global route optimization with local obstacle avoidance. The system employs convex optimization techniques for real-time trajectory generation while maintaining compliance with maritime regulations (COLREGs).

Software Architecture

ROS Framework: Robot Operating System as the backbone
Machine Learning: PyTorch/TensorFlow for AI models
Real-time Control: Extended Kalman Filter for state estimation
Communication: High-bandwidth primary + RF backup
Simulation: Gazebo for development and testing

Operational Modes

Flexible operation from manual control to full autonomy

🎮

Manual Mode

Direct operator control via high-reliability radio interface for development and complex scenarios

🤝

Semi-Autonomous

Mixed control with operator override capabilities and waypoint adjustment for supervised operation

🤖

Autonomous Mode

Full autonomous operation with route planning, waste collection, and charging decisions

Day-in-the-Life Scenario

Mission Start

Area Selection & Route Planning

TFV receives mission dispatch via operator prompt, scheduled cleaning, or autonomous weather-triggered cleaning detection.

Navigation

Autonomous Route Execution

Optimized route considering static obstacles, charging needs, waste disposal, and energy from solar panels.

Collection

Waste Detection & Harvesting

Local path planning for maximum coverage, waste detection, and conveyor belt collection with wildlife protection.

Return

Autonomous Return to Base

Return to docking station for charging and waste disposal, or safe anchoring in case of system failures.

Future Vision

Collaborative autonomous systems for comprehensive marine cleanup

Beyond WATER2030, the vision extends to collaborative fleets of autonomous vessels working together with supporting infrastructure for comprehensive marine environment restoration.

Collaborative System Components

🚢 TFV Fleet

Multiple vessels working in coordination
Shared mission planning and execution
Dynamic task allocation
Swarm intelligence behaviors

🏭 TFPod Infrastructure

Autonomous floating base stations
Solar panel arrays for energy
Hydrogen generation capabilities
Multi-vessel docking and charging

🚁 TFDrone Scouts

Aerial waste detection and mapping
Real-time ocean monitoring
Fleet coordination support
Extended range reconnaissance

Global Impact Potential

With successful development and deployment, WATER2030 technology could be adapted for various marine environments worldwide, from river systems to coastal areas and eventually open ocean applications. The open-source approach ensures global accessibility and collaborative improvement.