A system designed to clean litter from the seafloor is currently under development as part of a project aims to solve the problem of ocean litter with the help of unmanned vessels and artificial intelligence (AI).
The vessels and their related technology are being developed as part of the EU Horizon 2020-funded initiative known as the Search, Identification and Collection of marine Litter with Autonomous Robots (SeaClear) project. The unmanned vessels include a remotely operated vehicle (ROV) that will perform the actual cleaning underwater, another ROV for seabed scanning, an unmanned aerial vehicle (UAV) to provide overwatch and to conduct airborne detection of areas of pollution concentration, and a large unmanned surface vehicle (USV) equipped with a launch and recovery system (LARS), enabling it to function as a “mothership” for the ROVs and the UAV.
SeaClear came about after the project partners realised that the world’s oceans presently contain 26 million to 66 million tonnes of waste, with approximately 94 per cent of it located on the seafloor. The project aims to streamline the process of searching, identifying, and collecting marine litter using a team of autonomous robots that work collaboratively.
The SeaClear project team is a multi-national collaboration that includes the Delft University of Technology in the Netherlands, Hamburg-based maritime technology organisation Fraunhofer Centre for Maritime Logistics and Services (Fraunhofer CML), the Hamburg Port Authority, the Technical University of Munich, Marseilles-based engineering firm Subsea Tech (which also manufactured the USV and the ROVs), the Technical University of Cluj-Napoca in Romania, the Regional Development Agency Dubrovnik-Neretva County of Croatia, and the University of Dubrovnik (UniDu).
A multi-vehicle system
The USV first scans the area of interest of the sea bottom using a multibeam echosounder, which produces a 3D bathymetry map of the bottom. This serves as a reference map to which all other information about litter will be added. Some large litter like tyres or pipes may already be detectable from the bathymetry data, in which case this litter is already marked on the map.
As the USV also serves as the “mothership” of the system, all other robots deploy from and return to it. Robots communicate to the USV and get power from it via tethers. The computational resources required for sensing, control, and AI components are also hosted by the USV.
When the water is deemed to be sufficiently transparent, the UAV searches for litter from the air. The project team expects larger litter pockets to be identifiable in this way. Even in murky waters, the UAV still remains useful by scanning the surrounding area for obstacles that could otherwise hinder the scanning and cleanup phases.
A small ROV is deployed from the USV and performs close-up targeted scans of the sea bottom to find smaller litter. To this end, it uses a camera and a forward-looking sonar, together with possibly other sensors such as metal detectors. Identified litter is then placed on the reference map.
Litter is identitied with AI, deep-learning object recognition techniques. These deep networks are trained so as to differentiate litter from sea life and thereby ensure the system only collects what it should.
A larger variant of the observation ROV then deploys from the USV, goes to each piece of litter on the map, and grabs it with a gripper that is custom-made to interface with the unmanned vehicle. This gripper is equipped with a suction device that will help with picking up litter in difficult circumstances, such as when it is lying among plants. Each piece of litter is reacquired with high accuracy and then picked up.
To plan the paths and control the motion of both the observation and collection ROVs, SeaClear utilises intelligent techniques such as reinforcement learning and data-driven control.
Finally, a basket is deployed from the USV, and the collection ROV takes each piece of litter to deposit it in the basket for transportation to shore. The basket opening is specially designed to interface effectively with the gripper, and to prevent floating litter to escape back into the water.
SeaClear said the basket is not just a passive component, but actively sends signals to help the collection ROV localise itself relative to the opening.
Ongoing operational tests
Since the system consists of many components that must reliably work together under a multitude of different scenarios, SeaClear needed to make sure that the components will be able to accomplish their goal, hence the decision to undertake initial testing via computer simulation. The simulation is designed to run many different scenarios in a short time, and without the risk of damaging any of the equipment.
The first live tests of the SeaClear robots took place in September 2021 in the Dubrovnik area of Croatia. A collection of actual underwater litter was gathered in advance from a nearby location and was placed underwater in the test areas, attached to ropes and buoys for collection after the tests. The project team said this was necessary to introduce the robots to a representative and realistic litter collection. Video, sonar, and positioning datasets were also collected from unmanned aerial and underwater vehicles.
Two testing locations were chosen: 1) Lokrum island, where the robots were deployed via a mother ship operated by the UniDu, and 2) the mariculture laboratory of UniDu at Bistrina. The datasets collected will assist in the further development of localisation and litter detection algorithms.
The cleaning system, which works similarly to home robot vacuums, was able to see waste on the bottom of the sea and move towards it. A plastic bottle became the first item of litter to be picked up from the seabed.
The second round of trials of the took place in the Port of Hamburg between May 10 and 12, 2022. The port of Hamburg is the second test site of the project, offering completely different conditions and challenges compared to the first test site in Dubrovnik. Underwater visibility is considerably lower, often lower than a few centimetres, while there is heavy traffic present from commercial ships. SeaClear said this makes the use of vision cameras for detecting the litter unusable, with the only reliable detection alternative being sonar sensors.
The SeaClear team’s goal is to make the system commercially viable. The team aims to exploit the results of the project both by selling the system and by providing it as a service to local authorities, ports, or other stakeholders.
The project team is set to meet again in Marseilles this September. The project itself will span four years with completion scheduled for December 31, 2023.
The best maritime site on the web. The sea's our scene!