The oceans currently contain between 26 and 66 million tons of waste, with 94% deposited on the seabed. Until now, collection efforts have focused primarily on surface debris, and those undertaken to remove underwater debris involve human divers working in hazardous conditions. A group of researchers, supported by the European Union (EU), is developing the first fleet of intelligent autonomous robots to combat debris both on the surface and on the ocean floor.
The EU-funded SeaClear2.0 project develops the first autonomous robotic and artificial intelligence system capable of finding and removing debris from the ocean floor.
The idea is to deploy a fleet of intelligent robots to first identify and then remove marine debris. The system uses a combination of aerial drones, underwater rovers, autonomous surface ships and robotic grippers custom made.
Wearing artificial intelligencerobots can identify, locate and collect garbage autonomously. The team aims to significantly expand the capabilities of the predecessor system, SeaClear1.0, which is already passing tests with flying colors.
The technology core of SeaClear2.0 is a much-improved robotic system for automated garbage collection, which can go deeper than SeaqClear1.0, lift heavier garbage, and tackle surface garbage as well.
SeaClear2.0 will detect debris with a high resolution sonaras well as other types of sensors. It will also develop intelligent and maneuverable claw that can pick up large trash, such as tires, bicycles or pipes.
The first tests, in the Mediterranean
To transport bulkier and heavier trash, the scientists will design a new tender with flexible dividing chambers for different fractions of trash, similar to adjustable drawer dividers.
LTrash from the surface will be collected by autonomous mobile equipment that can work in confined and shallow areas.
With almost 500 million tons of plastics that reach European seas every yeartechnological solutions to identify and eliminate garbage must be combined with “social interventions to prevent and reduce the production and release of garbage by citizens“, say the promoters of the project.
Thus, SeaClear2.0 provides empower and activate citizens through engagement activities including geographic storytelling, a gamified trash reporting app, cleanups, exhibits, contests, and art installations.
The team will also propose best sorting and recycling solutions to increase the value of collected trash, and will work with local communities to develop recommendations for novel policies and thus address existing gaps in current policies.
The new improved system will be shown in three full-scale demonstrations and three pilot tests around the Mediterranean. In addition, the project will finance five subprojects to validate parts of the system and the approach in other regions of the Mediterranean basin.
The SeaClear2.0 project started on January 1st and will last four years. The launch meeting is already planned for February 14, in Delft, the Netherlands.
A ‘mothership’, a drone and two ‘turtles’
The process begins with the use of an unmanned surface craft, the ‘SeaCat’, which scans the seabedmake a 3D bathymetric map and mark any large debris you find.
The ‘SeaCat’ also serves as the ‘mothership‘ system: all other robots deploy, communicate and return to it. They also get power from it through wires. The computational resources necessary for the detection, control and artificial intelligence components are also housed in ‘SeaCat’.
When the water is clear enough, a Unmanned aerial vehicle (a drone DJI Matrice M210 RTK Modified) look for garbage from the air. Larger garbage bags are expected to be identifiable in this way and will inform the more detailed search using the underwater robot in the next step.
A small unmanned underwater vehicle, the ‘miniTurtle’ is then deployed from the ‘SeaCat’ and performs specific close-up explorations of the seabed to find smaller trash. To this end, it uses a forward-looking camera and sonar, along with other sensors, such as metal detectors. All this is added to the map that locates the waste.
Trash is identified using artificial intelligence and deep learning object recognition techniques. These deep networks are trained to differentiate litter from marine life and therefore ensure that the system only collects what it should.
then act for a second marine robot, named ‘Turtle‘, equipped with a clamp that includes a suction device that helps to pick up trash in difficult circumstances, as when it is among algae. Each piece of trash is recovered with great precision and then collected.
A budget of 9 million
To plan the routes and control the movement of the observation and collection robots, they are exploited clever techniques, such as reinforcement learning and data-driven control.
Finally, it unfolds a basket from the ‘SeaCat’, and the robot ‘Turtle’ takes each piece of garbage and deposits it in it to then transport it to the shore. The basket opening is specially designed to effectively interact with the grapple to prevent floating debris from returning to the water.
But the basket is not just a passive component, it actively send signals to help the robot locate itself in the perfect place to take advantage of the opening.
The project receives funding from the EU’s ‘Horizon Europe’ programme. Its total budget is just over 9 million eurosof which almost 8 come from community funds.
The consortium is made up of 13 partners from 9 countries, with a mix of expertise in public engagement, policy making, robotic perception and control, artificial intelligence, marine and dive technology and operations, and garbage sorting and recycling.
SeaClear2.0 (Scalable Full-cycle Marine Litter Remediation in the Mediterranean: Robotic and Participatory Solutions) will end in December 2026.
‘SeaClear’ project website: https://seaclear-project.eu/
Contact of the Environment section: [email protected]