What is your SeaClear use case?

Coastal Cities and Municipalities

  • Challenge: Struggle with the impact of marine litter on natural beauty, tourism, and local economy.
  • SeaClear Solution: Offers efficient, autonomous litter collection, enhancing beach and sea cleanliness, boosting tourism appeal, and supporting sustainable ocean initiatives.
  • Impact: Cleaner coastal environments, improved tourist attraction, and healthier marine ecosystems.

Ports and Harbors

  • Challenge: Debris accumulation posing risks to safe and efficient maritime operations.
  • SeaClear Solution: Efficient underwater litter removal ensures safer harbor operations and navigational safety, contributing to eco-friendly port environments.
  • Impact: Enhanced operational safety, reduced navigational hazards, and promotion of sustainable port activities.

Tourism and Hospitality Sector

  • Challenge: Marine litter detracts from the attractiveness and safety of popular tourist destinations.
  • SeaClear Solution: Maintains the pristine condition of beaches and underwater areas, aligning with eco-tourism and sustainable practices.
  • Impact: Enhanced tourist appeal, preserved natural beauty, and support for eco-friendly tourism growth.

Marine Conservation and Research

  • Challenge: Need for effective methods to protect marine biodiversity and research opportunities.
  • SeaClear Solution: Aids in marine conservation by removing harmful waste and providing data for research, while minimizing disturbance to marine life.
  • Impact: Protection of marine habitats, support for biodiversity, and valuable contributions to marine research.

Download our pitch deck

 You can download our pitch deck using this link

Book a meeting

Please use our contact form for booking a meeting with us

Read more: Book a meeting with SeaClear

AI and Robots Revolutionize Marine Litter Collection with SeaClear Project

European researchers created a functional team of smart robots designed to clean litter from the seafloor. The SeaClear system, created within a four-year European research project which ended in December 2023, has successfully passed a series of tests in both clear and murky waters.

The SeaClear system is composed of several interconnected components. The base vessel, SeaCAT, acts as the central unit, deploying and managing two underwater remotely operated vehicles (ROVs): the Mini Tortuga for exploration and the larger, Tortuga ROV, for litter collection. The litter is deposited in a special basket. Additionally, an aerial drone monitors and maps the sea surface, assisting in the identification of litter hotspots. These elements work in concert to create a map of the ocean floor, detect litter, and subsequently collect and remove it efficiently.

In its current stage, the SeaClear system can lift up to 7 kg, the gripper can fit the volume of two 2L soda bottles, and the robots collect litter in waters up to tens of meters deep. When improved for commercial operation, the system will work with a 70% smaller cost than divers.

The research was funded by a Horizon 2020 project led by TU Delft in the Netherlands, concluding in December 2023. "By the project’s end, we had a fully operational system demonstrating its functionality," said the project lead. This groundbreaking innovation showcases the potential of autonomous robotics to address the pressing issue of ocean litter.

The issue at the heart of the SeaClear project is the escalating crisis of ocean litter, a problem that poses a severe threat to marine ecosystems, wildlife, and human health. Oceans are currently burdened with an estimated 26 to 66 million tons of waste, predominantly consisting of plastic. "When we look at the sea surface, we only see a beautiful blue area. But once you get under the surface, you see the ugly reality", says Iva Pozniak from the Regional Development Agency Dubrovnik-Neretva County in Croatia. More than 90% of the waste  is not visible but rather lies at the bottom of the sea.

This waste not only degrades the beauty of marine landscapes, but it also destroys habitats, entangles and poisons marine life, and disrupts entire underwater ecosystems. Microplastics, which are produced by the breakdown of larger plastic items, enter the food chain and pose health risks to both marine life and humans. The problem is made worse by the fact that a large amount of this litter ends up on the seafloor, where it is difficult to find and remove. SeaClear's mission is to combat this underwater litter problem using innovative, autonomous robotic technology, aiming to clean our oceans and protect their biodiversity.

"Nothing in nature helps clean the oceans. We must innovate solutions ourselves," says Stefan Sonowski of TU Munchen in Germany. While a solution does not yet exist in nature, he and his colleagues did draw inspiration from nature, particularly in designing the gripper's honeycomb structure, which is lightweight yet sturdy and allows small marine creatures to escape while retaining litter.

One of the key features of the SeaClear system is its adaptability to different conditions, including varied water properties and different kinds of litter. “Unfortunately, litter is everywhere, but different kinds of litter occur in different places," explains Cosmin Delea from Fraunhofer CML, Germany. For example, Hamburg Port, one of the test sites of SeaClear, had heavier industrial waste, while in the tourist areas of Croatia there is lighter, more diverse waste such as plastic bottles and bags.

Artificial intelligence (AI) is pivotal to SeaClear’s functionality. AI algorithms accurately detect and identify litter, and have been trained to differentiate it from marine life. Integrating machine learning and computer vision lets the underwater robots navigate underwater environments and make decisions.

The newly emerged large-language models like ChatGPT might help further with high-level planning: determining the optimal actions for the system. However, difficult challenges in robotic automation still remain at the lower level. “Imagine asking a robot to bring you a beer from the fridge," explains professor Lucian Bușoniu from the Technical University of Cluj-Napoca, Romania. “How to orient the gripper, grip the bottle, apply the right forces — this interface between robot and the real world is where things get messy.” For such tasks in SeaClear, solutions will continue to involve engineering and testing.

Machine learning also helps guide components to prevent cable tangling. Some components connect to the SeaCat mothership via cables, enabling them to operate underwater for hours. “We developed an algorithm controlling cable slack,” says Ivana Palunko of the University of Dubrovnik, stressing that “Cable management is very important underwater.”

As SeaClear might become a feasible solution for cleaning the ocean bed, the team must also consider the cross-cutting legal implications and impact on other actors involved in maritime business. “SeaClear is technically autonomous and can do the job by itself, but in reality we need to consider commercial and recreational shipping, kayaks, industrial ships, and divers that are also in the water," warns Claudia Hertel-Ten Eikelder from the Hamburg Port Authority, where several tests have taken place. 

With proof-of-concept experiments demonstrated, SeaClear could now be used commercially, says Yves Chardard of SeaClear’s French partner SubSea Tech. This solution could be attractive to entities such as ports, tourist sites, environmental institutions, and other civil and military institutions.

For the development of the SeaClear system, in the subsequent project SeaClear 2.0, “we need to go faster, heavier, and deeper," says Chardard. “With SeaClear 1.0, we are picking up plastic bottles, shoes and other similar waste. With SeaClear 2.0, we want to pick up tires, washing machines, and in general heavier things. It’s much more difficult to lift 250 kilos from the seabed, so we want to focus on that.”

Building on SeaClear’s success, the team has secured 9 million euros in Horizon Europe funding and cofunding for SeaClear 2.0, to build a more robust system for surface and deep Mediterranean litter. In addition to robotics, this followup project also focuses on community engagement and policy, with activities like gamified apps and educational programs to curb litter production. With demonstrations planned across the Mediterranean, the project’s diverse 13-partner, 9-country consortium combines public engagement, AI, marine technology and recycling expertise, enabling a comprehensive approach to conservation. The SeaClear2.0 project started in January 2023 and is a part of the EU Mission "Restore our Ocean and Waters”; follow our social media at https://linktr.ee/seaclear2.0 to find out more.

Videos:

  • Final demonstration in Dubrovnik
  • SeaClear Demonstration in Hamburg:
     
  • Concept video for follow-up project SeaClear2.0:

Romanian version of the press release

Tags: press release

Read more: SeaClear: The Robots are Ready to Clean the Ocean

Lesson plan about Keeping the Oceans Clean

What's Inside the Lesson Plan?

A complete, 50 minutes school lesson about ocean pollution

This is appropriate for children between 12–15 years old

Includes information about

  • Sources of pollution
  • Human health effects
  • Environment effects
  • Self-reflection questions
  • SeaClear’s initiative of cleaning the litter using robots
  • Quiz / Recap questions
  • Interesting graphics
  • An ongoing online challenge

And guess what: You will receive a certificate!

If you present the lesson in your classroom, or you invite a SeaClear member to present the lesson, you will receive a SeaClear certificate as a #SeaChampion!

You can download the lesson plan

Or

Contact us to have a presentation in your school

Infographic about ocean pollution

This visually engaging tool that simplifies the complex topic of Ocean Pollution

It summarizes information about:

  • Sources of pollution
  • Types of pollution
  • Effects on the ocean sea life
  • The SeaClear robotic system for cleaning the ocean
  • Actions to take

The infographic can be printed, or it can be used in online projects. It is appropriate for 10–16 years old pupils. You can download it using this link

Tags: media

Read more: Educational resources

After the success of the first autonomous robotic system designed to search for, identify, and collect seafloor litter, a consortium of European researchers will develop a team of robots to collect litter from the surface and deeper regions of the Mediterranean. The teams responsible for the Horizon 2020 SeaClear project have announced that they have expanded their consortium and received funding to begin follow-up work to further develop the robotic system and holistically address the issue of marine litter through the project: SeaClear2.0  – Scalable Full-cycle Marine Litter Remediation in the Mediterranean: Robotic and Participatory Solutions. SeaClear2.0 will go beyond technological innovation, integrating cutting-edge technologies into a comprehensive approach that engages communities in finding solutions to marine litter pollution, increases waste value, and contributes to science-based policy making.

SeaClear 2.0, like its predecessor, attempts to use robots to combat one of the world's most serious environmental issues: marine litter. Cleaning it up using divers poses danger to human lives and is economically overwhelming. 

The new EU-funded initiative will deploy a fleet of smart robots to identify and remove marine debris, building on the success of SeaClear 1.0. The system employs a combination of aerial drones, underwater rovers, autonomous surface ships, and custom-built robotic grippers to do this. Using artificial intelligence, robots can autonomously identify, locate, and collect litter on their own. The team is aiming to significantly expand the capabilities of the SeaClear1.0 system (shown on the right of the concept figure), which is already passing tests with flying colours.

The technological core of SeaClear2.0 is a heavily upgraded robotic system for automated litter collection, which can go deeper, lift heavier litter, and tackle surface litter (see the concept figure, left and centre). The project starts by improving sensing with high-resolution sonar, as well as other types of sensors like microplastic and electromagnetic sensors. A smart, maneuverable grapple will be developed that can pick-up larger litter, such as tires, bicycles, or pipes. To carry the bulkier, heavier litter, the scientists will design a new tender with flexible partitioning chambers for different litter fractions, similar to adjustable drawer dividers. Surface litter will be collected via an autonomous mobile team that can work in confined and shallow areas. This new system will be able to work in tandem with the SeaClear1.0 version.

With nearly half a billion tonnes of plastics entering European seas every year, technological solutions to identify and remove litter must be coupled with social interventions to prevent and reduce the production and release of litter by citizens. SeaClear2.0 will empower and activate citizens through engagement activities including geographical storytelling, a gamified litter-reporting app., clean-ups, exhibitions, competitions, and artistic installations. The team will also  propose better sorting and recycling solutions to increase the value of the litter collected, and will work with local communities to develop novel policy recommendations to address policy gaps. 

The new, improved system will be demonstrated in 3 full-scale demonstrations and 3 pilot tests throughout the Mediterranean. In addition, the project will fund 5 subprojects to validate parts of the system and approach in other regions of the Mediterranean basin.

The SeaClear2.0 project will start on January 1st 2023, and will run for four years. The kick-off meeting has already been planned to take place in Delft, the Netherlands on the 14th of February, 2023. This project receives funding from the European Union’s Horizon Europe programme under grant agreement No. 101093822. The total budget of the project is €9,086,305.00, out of which €7,971,863.50 is funded by the EU. The consortium consists of 13 partners from 9 countries, with a blend of expertise in public engagement, policy-making, robotic perception and control, artificial intelligence, marine, and diving technology and operations, and litter sorting and recycling: Delft University of Technology (Netherlands, project coordinator), Regional Agency Dunea (Croatia), Fraunhofer (Germany); Hamburg Port Authority (Germany), Isotech (Cyprus), M.Danchor (Israel), Subsea Tech (France), Técnicas y Obras Subacuáticas (TECNOSUB) (Spain), Technische Universitaet Muenchen (Germany), University of Dubrovnik (Croatia), Technical University of Cluj-Napoca (Romania), Veolia (France), and Venice Lagoon Plastic Free (Italy).

Read more: SeaClear2.0 - An autonomous robotic system...

Learning for multi-robot and networked systems

Organisers

  • Bart De Schutter, Delft University of Technology (This email address is being protected from spambots. You need JavaScript enabled to view it.)
  • Lucian Busoniu, University of Cluj-Napoca (This email address is being protected from spambots. You need JavaScript enabled to view it.)
  • Stefan Sosnowski, Technical University of Munich (This email address is being protected from spambots. You need JavaScript enabled to view it.)

Abstract

Multi-robot systems involve two or more autonomous robots that are working together to achieve one or more well-defined objectives. Individual robots may be rather simple and by themselves unable to achieve the desired goals. However, the real power – and at the same time also the major challenge – lies in the cooperation and coordination of the individual robots so as to jointly achieve the specified objectives. Multi-robot interaction is moreover subject to challenges stemming from the networked and communication structure of the system. Although multi-robot systems have attracted significant attention worldwide, research in this area is still in its infancy. This open invited track aims at bringing together contributions covering the broad area of computational-intelligence, machine-learning, networked-control and AI-based methods for multi-agent decision-making in multi-robot systems as well as for coordination in networked systems in general. In addition to papers proposing new fundamental results, we also explicitly solicit papers that show the potential of multi-robot interaction and coordination in networked systems in experimental set-ups and real-life applications. Authoritative survey papers are also welcome.

Paper submission

The initial paper submission deadline is the 31st of October 2022. To submit a paper to this track, go to the IFAC PaperPlaza website, search for the IFAC World Congress, and go through the submission procedure choosing one of the categories: Open invited track survey paper, Open invited track paper, or Open invited track discussion paper. You will need to enter the track code "5fic1".

Details and topics

This open invited session aims to:

  • bring together novel results in computational-intelligence, machine-learning, networked-control and AI-based methods for multi-agent decision-making for multi-robot systems as well as for coordination in networked systems in general,
  • address emerging challenges in decision-making for multi-robot systems and coordination in networked systems, including dealing with highly uncertain or time-varying environments, restricted computational and communication resources, non-conventional environments (e.g. underwater multi-robot systems), etc.
  • present emerging relevant applications.

We encourage the submission of research on multi-robot and networked systems related, but not limited to:

  • computational intelligence for multi-agent systems
  • robust decision-making methods for multi-robot systems
  • hybrid computational intelligence techniques
  • reinforcement learning
  • transfer learning
  • learning for prediction and control
  • distributed and federated learning
  • adaptive dynamic programming for multi-agent systems
  • networked systems and cooperative control
  • game theory for multi-robot interaction
  • neuro-fuzzy and deep learning approaches
  • applications and demonstrations

 

Read more: Proposal for open invited track for IFAC World...

Future events

Past events

Read more: Events

(Translated from original press release)

Eight partners accross Europe: Hamburg Port Authority (HPA), Fraunhofer Center for Maritime Logistics and Services (CML), Delft University of Technology (TUD), Dubrovnik Regional Development Company (DUNEA), SUBSEA TECH SAS (SST), Technical University of Cluj-Napoca (TUC),  Technical University of Munich (TUM), and the University of Dubrovnik (UNIDU) are working on the EU-funded project SEACLEAR on the development of an autonomous system to clean the seabed.

The SEACLEAR system consists of an autonomous vessel with two underwater robots that identify and collect underwater debris. The system, whose individual components have been developed by various project partners since January 2020, was presented at the ITS Congress 2 021. Now the individual components are being brought together for the first time and the processes are being tested under real conditions as part of a test campaign. The area of the Hansahafen in the Port of Hamburg acts as a test location.

SeaClear concept 2 small

“Although we in Hamburg do not have the challenge of having to salvage large amounts of plastic waste from the Elbe, the port offers ideal conditions for the tests of the SEACLEAR system. The turbid water of the Elbe, the tidal currents and also the ship traffic pose special challenges for the system. For us, these tests also provide important experience when using autonomous systems says Jens Meier, CEO of the Hamburg Port Authority.

“We are pleased to be testing the system for the first time in the port of Hamburg,” says Carlos Jahn, head of the Fraunhofer Center for Maritime Logistics and Services. “The results of the week-long test runs will drive our developments in the area of system integration as the project progresses.”

In the course of the tests, among other things, prepared samples of waste particles are lowered to the bottom of the water. The task of the small diving robot is to record and mark the finds in the previously created map of the area.

Tags: press release

Read more: Fighting plastic waste: SEACLEAR tests in the...

Journal Publications

 

  1. Jerome Weston, Domagoj Tolić and Ivana Palunko. Application of Hamilton–Jacobi–Bellman Equation/Pontryagin's Principle for Constrained Optimal Control. Journal of Optimization Theory and Applications, January 2024. URL, DOI BibTeX

    @article{Weston2024,
    	author = "Weston, Jerome and Toli{\'{c}}, Domagoj and Palunko, Ivana",
    	title = "Application of Hamilton--Jacobi--Bellman Equation/Pontryagin's Principle for Constrained Optimal Control",
    	journal = "Journal of Optimization Theory and Applications",
    	year = 2024,
    	month = "Jan",
    	day = 09,
    	abstract = "This article applies novel results for infinite- and finite-horizon optimal control problems with nonlinear dynamics and constraints. We use the Valentine transformation to convert a constrained optimal control problem into an unconstrained one and show uniqueness of the value function to the corresponding Hamilton--Jacobi--Bellman (HJB) equation. From there, we show how to approximate the solution of the initial (in)finite-horizon problem with a family of solutions that is {\$}{\$}{\backslash}varGamma {\$}{\$}-convergent. Optimal solutions are efficiently obtained via a solver based on Pontryagin's Principle (PP). The proposed methodology is demonstrated on the path planning problem using the full nonlinear dynamics of an unmanned aerial vehicle (UAV) and autonomous underwater vehicle (AUV) involving state constraints in 3D environments with obstacles.",
    	issn = "1573-2878",
    	doi = "10.1007/s10957-023-02364-4",
    	url = "https://doi.org/10.1007/s10957-023-02364-4"
    }
    
  2. Benjamin Kelenyi and Levente Tamas. D3GATTEN: Dense 3D Geometric Features Extraction and Pose Estimation Using Self-Attention. IEEE Access 11():7947-7958, 2023. DOI BibTeX

    @article{10024269,
    	author = "Kelenyi, Benjamin and Tamas, Levente",
    	journal = "IEEE Access",
    	title = "D3GATTEN: Dense 3D Geometric Features Extraction and Pose Estimation Using Self-Attention",
    	year = 2023,
    	volume = 11,
    	number = "",
    	pages = "7947-7958",
    	doi = "10.1109/ACCESS.2023.3238901"
    }
    
  3. Ioana Lal, Irinel-Constantin Morarescu, Jamal Daafouz and Lucian Busoniu. Optimistic planning for control of hybrid-input nonlinear systems. Automatica 154:111097, 2023. DOI BibTeX

    @article{aut23-ioana,
    	title = "Optimistic planning for control of hybrid-input nonlinear systems",
    	author = "Lal, Ioana and Morarescu, Irinel-Constantin and Daafouz, Jamal and Busoniu, Lucian",
    	journal = "Automatica",
    	volume = 154,
    	pages = 111097,
    	year = 2023,
    	doi = "https://doi.org/10.1016/j.automatica.2023.111097"
    }
    
  4. Ivica Nakić, Domagoj Tolić, Zoran Tomljanović and Ivana Palunko. Numerically efficient H∞ analysis of cooperative multi-agent systems. Journal of the Franklin Institute 359(16):9110-9128, 2022. URL, DOI BibTeX

    @article{NAKIC20229110,
    	title = "Numerically efficient H∞ analysis of cooperative multi-agent systems",
    	journal = "Journal of the Franklin Institute",
    	volume = 359,
    	number = 16,
    	pages = "9110-9128",
    	year = 2022,
    	issn = "0016-0032",
    	doi = "https://doi.org/10.1016/j.jfranklin.2022.09.013",
    	url = "https://www.sciencedirect.com/science/article/pii/S0016003222006561",
    	author = "Ivica Nakić and Domagoj Tolić and Zoran Tomljanović and Ivana Palunko",
    	abstract = "This article proposes a numerically efficient approach for computing the maximal (or minimal) impact one agent has on the cooperative system it belongs to. For example, if one is able to disturb/bolster merely one agent in order to maximally disturb/bolster the entire team, which agent to choose? We quantify the agent-to-system impact in terms of H∞ norm whereas output synchronization is taken as the underlying cooperative control scheme. The agent dynamics are homogeneous, second order and linear whilst communication graphs are weighted and undirected. We devise simple sufficient conditions on agent dynamics, topology and output synchronization parameters rendering all agent-to-system H∞ norms to attain their maxima in the origin (that is, when constant disturbances are applied). Essentially, we quickly identify bottlenecks and weak/strong spots in multi-agent systems without resorting to intense computations, which becomes even more important as the number of agents grows. Our analyses also provide directions towards improving communication graph design and tuning/selecting cooperative control mechanisms. Lastly, numerical examples with a large number of agents and experimental verification employing off-the-shelf nano quadrotors are provided."
    }
    
  5. Thomas Beckers, Leonardo J Colombo and Sandra Hirche. Safe trajectory tracking for underactuated vehicles with partially unknown dynamics. Journal of Geometric Mechanics 14(4):491-505, 2022. URL, DOI BibTeX

    @article{thomas_Beckers_2022,
    	title = "Safe trajectory tracking for underactuated vehicles with partially unknown dynamics",
    	journal = "Journal of Geometric Mechanics",
    	volume = 14,
    	number = 4,
    	pages = "491-505",
    	year = 2022,
    	issn = "1941-4889",
    	doi = "10.3934/jgm.2022018",
    	url = "https://www.aimsciences.org/article/doi/10.3934/jgm.2022018",
    	author = "Thomas Beckers and Leonardo J. Colombo and Sandra Hirche",
    	keywords = "Tracking control, underactuated systems, learning, data-driven methods, safe control, formal methods"
    }
    
  6. Tudor Santejudean and Lucian Busoniu. Online learning control for path-aware global optimization with nonlinear mobile robots. Control Engineering Practice 126:105228, 2022. URL, DOI BibTeX

    @article{santejudean_online_2022,
    	author = "Santejudean, Tudor and Busoniu, Lucian",
    	title = "Online learning control for path-aware global optimization with nonlinear mobile robots",
    	journal = "Control Engineering Practice",
    	year = 2022,
    	note = "In press",
    	volume = 126,
    	pages = 105228,
    	doi = "https://doi.org/10.1016/j.conengprac.2022.105228",
    	url = "http://busoniu.net/files/papers/cep22-tudor.pdf"
    }
    
  7. Matthias Rosynski and Lucian Buşoniu. A Simulator and First Reinforcement Learning Results for Underwater Mapping. Sensors 22(14):5384, 2022. URL, DOI BibTeX

    @article{rosynski_simulator_2022,
    	author = "Matthias Rosynski and Lucian Bu\c{s}oniu",
    	title = "A Simulator and First Reinforcement Learning Results for Underwater Mapping",
    	journal = "Sensors",
    	year = 2022,
    	volume = 22,
    	url = "https://www.mdpi.com/1424-8220/22/14/5384/pdf?version=1658453246",
    	number = 14,
    	pages = 5384,
    	doi = "https://doi.org/10.3390/s22145384"
    }
    
  8. Petar Bevanda, Max Beier, Sebastian Kerz, Armin Lederer, Stefan Sosnowski and Sandra Hirche. Diffeomorphically Learning Stable Koopman Operators. IEEE Control Systems Letters 6():3427-3432, 2022. URL, DOI BibTeX

    @article{bevanda_diffeomorphically_2022,
    	author = "Bevanda, Petar and Beier, Max and Kerz, Sebastian and Lederer, Armin and Sosnowski, Stefan and Hirche, Sandra",
    	journal = "IEEE Control Systems Letters",
    	title = "Diffeomorphically Learning Stable Koopman Operators",
    	year = 2022,
    	volume = 6,
    	number = "",
    	pages = "3427-3432",
    	doi = "https://doi.org/10.1109/LCSYS.2022.3184927",
    	url = "https://mediatum.ub.tum.de/doc/1636838/3ptpmd2tuvg80a6m608tfg1b0.KoopFlows_lcssRev_arxiv_.pdf"
    }
    
  9. Armin Lederer, Zewen Yang, Junjie Jiao and Sandra Hirche. Cooperative Control of Uncertain Multiagent Systems via Distributed Gaussian Processes. IEEE Transactions on Automatic Control 68(5):3091-3098, 2023. DOI BibTeX

    @article{9882336,
    	author = "Lederer, Armin and Yang, Zewen and Jiao, Junjie and Hirche, Sandra",
    	journal = "IEEE Transactions on Automatic Control",
    	title = "Cooperative Control of Uncertain Multiagent Systems via Distributed Gaussian Processes",
    	year = 2023,
    	volume = 68,
    	number = 5,
    	pages = "3091-3098",
    	doi = "10.1109/TAC.2022.3205424"
    }
    
  10. Athina Ilioudi, Azita Dabiri, Ben J Wolf and Bart De Schutter. Deep Learning for Object Detection and Segmentation in Videos: Toward an Integration With Domain Knowledge. IEEE Access, 2022. URL, DOI BibTeX

    @article{ilioudi_Deep_2022,
    	title = "Deep {Learning} for {Object} {Detection} and {Segmentation} in {Videos}: {Toward} an {Integration} {With} {Domain} {Knowledge}",
    	url = "https://pure.tudelft.nl/ws/portalfiles/portal/116826331/Deep_Learning_for_Object_Detection_and_Segmentation_in_Videos_Toward_an_Integration_With_Domain_Knowledge.pdf",
    	doi = "10.1109/ACCESS.2022.3162827",
    	abstract = "Deep learning has enabled the rapid expansion of computer vision tasks from image frames to video segments. This paper focuses on the review of the latest research in the field of computer vision tasks in general and on object localization and identification of their associated pixels in video frames in particular. After performing a systematic analysis of the existing methods, the challenges related to computer vision tasks are presented. In order to address the existing challenges, a hybrid framework is proposed, where deep learning methods are coupled with domain knowledge. An additional feature of this survey is that a review of the currently existing approaches integrating domain knowledge with deep learning techniques is presented. Finally, some conclusions on the implementation of hybrid architectures to perform computer vision tasks are discussed.",
    	urldate = "2022-05-05",
    	journal = "IEEE Access",
    	author = "Ilioudi, Athina and Dabiri, Azita and Wolf, Ben J. and De Schutter, Bart",
    	month = "",
    	year = 2022
    }
    
  11. Daniël M Bot, Ben J Wolf and Sietse M Netten. The Quadrature Method: A Novel Dipole Localisation Algorithm for Artificial Lateral Lines Compared to State of the Art. Sensors 21(13):4558, 2021. URL, DOI BibTeX

    @article{bot_Quadrature_2021,
    	title = "The {Quadrature} {Method}: {A} {Novel} {Dipole} {Localisation} {Algorithm} for {Artificial} {Lateral} {Lines} {Compared} to {State} of the {Art}",
    	volume = 21,
    	copyright = "http://creativecommons.org/licenses/by/3.0/",
    	issn = "1424-8220",
    	shorttitle = "The {Quadrature} {Method}",
    	url = "https://www.mdpi.com/1424-8220/21/13/4558/pdf?version=1625722411",
    	doi = "10.3390/s21134558",
    	abstract = "The lateral line organ of fish has inspired engineers to develop flow sensor arrays—dubbed artificial lateral lines (ALLs)—capable of detecting near-field hydrodynamic events for obstacle avoidance and object detection. In this paper, we present a comprehensive review and comparison of ten localisation algorithms for ALLs. Differences in the studied domain, sensor sensitivity axes, and available data prevent a fair comparison between these algorithms from their original works. We compare them with our novel quadrature method (QM), which is based on a geometric property specific to 2D-sensitive ALLs. We show how the area in which each algorithm can accurately determine the position and orientation of a simulated dipole source is affected by (1) the amount of training and optimisation data, and (2) the sensitivity axes of the sensors. Overall, we find that each algorithm benefits from 2D-sensitive sensors, with alternating sensitivity axes as the second-best configuration. From the machine learning approaches, an MLP required an impractically large training set to approach the optimisation-based algorithms’ performance. Regardless of the data set size, QM performs best with both a large area for accurate predictions and a small tail of large errors.",
    	language = "en",
    	number = 13,
    	urldate = "2022-05-05",
    	journal = "Sensors",
    	author = "Bot, Daniël M. and Wolf, Ben J. and van Netten, Sietse M.",
    	month = "",
    	year = 2021,
    	note = "Number: 13 Publisher: Multidisciplinary Digital Publishing Institute",
    	keywords = "artificial lateral line, dipole localisation, hydrodynamic imaging, neural networks",
    	pages = 4558
    }
    
  12. Petar Bevanda, Stefan Sosnowski and Sandra Hirche. Koopman operator dynamical models: Learning, analysis and control. Annual Reviews in Control 52:197-212, 2021. URL, DOI BibTeX

    @article{bevanda_koopman_2022,
    	title = "Koopman operator dynamical models: Learning, analysis and control",
    	journal = "Annual Reviews in Control",
    	volume = 52,
    	pages = "197-212",
    	year = 2021,
    	issn = "1367-5788",
    	doi = "https://doi.org/10.1016/j.arcontrol.2021.09.002",
    	url = "https://arxiv.org/pdf/2102.02522.pdf",
    	author = "Petar Bevanda and Stefan Sosnowski and Sandra Hirche",
    	keywords = "Koopman operator, Dynamical models, Representation learning, System analysis, Data-based control",
    	abstract = "The Koopman operator allows for handling nonlinear systems through a globally linear representation. In general, the operator is infinite-dimensional – necessitating finite approximations – for which there is no overarching framework. Although there are principled ways of learning such finite approximations, they are in many instances overlooked in favor of, often ill-posed and unstructured methods. Also, Koopman operator theory has long-standing connections to known system-theoretic and dynamical system notions that are not universally recognized. Given the former and latter realities, this work aims to bridge the gap between various concepts regarding both theory and tractable realizations. Firstly, we review data-driven representations (both unstructured and structured) for Koopman operator dynamical models, categorizing various existing methodologies and highlighting their differences. Furthermore, we provide concise insight into the paradigm’s relation to system-theoretic notions and analyze the prospect of using the paradigm for modeling control systems. Additionally, we outline the current challenges and comment on future perspectives."
    }
    

Conference Publications

 

  1. Athina Ilioudi, Ben J Wolf, Azita Dabiri and Bart De Schutter. Towards establishing an automated selection framework for underwater image enhancement methods. In OCEANS 2023 - Limerick (). 2023, 1-6. DOI BibTeX

    @inproceedings{10244710,
    	author = "Ilioudi, Athina and Wolf, Ben J. and Dabiri, Azita and De Schutter, Bart",
    	booktitle = "OCEANS 2023 - Limerick",
    	title = "Towards establishing an automated selection framework for underwater image enhancement methods",
    	year = 2023,
    	volume = "",
    	number = "",
    	pages = "1-6",
    	doi = "10.1109/OCEANSLimerick52467.2023.10244710"
    }
    
  2. Matthias Rosynski, Alexandru Pop and Lucian Busoniu. Active search and coverage using point-cloud reinforcement learning. In Proceedings of the 27th International Conference on System Theory, Control and Computing. 2023. BibTeX

    @inproceedings{icstcc23,
    	author = "Matthias Rosynski and Alexandru Pop and Lucian Busoniu",
    	title = "Active search and coverage using point-cloud reinforcement learning",
    	booktitle = "Proceedings of the 27th International Conference on System Theory, Control and Computing",
    	year = 2023,
    	date = "2023-10-11",
    	address = "Timisoara, Romania"
    }
    
  3. Petar Bevanda, Max Beier, Armin Lederer, Stefan Sosnowski, Eyke Hüllermeier and Sandra Hirche. Koopman Kernel Regression. In Advances in Neural Information Processing Systems 37. 2023. DOI BibTeX

    @inproceedings{KKR_neurips2023,
    	title = "Koopman Kernel Regression",
    	author = {Bevanda, Petar and Beier, Max and Lederer, Armin and Sosnowski, Stefan and H{\"u}llermeier, Eyke and Hirche, Sandra},
    	booktitle = "Advances in Neural Information Processing Systems",
    	volume = 37,
    	year = 2023,
    	eprint = "2305.16215",
    	doi = "https://doi.org/10.48550/arXiv.2305.16215"
    }
    
  4. Bilal Yousuf, Zsofia Lendek and Lucian Buşoniu. Multi-Agent Exploration-Based Search for an Unknown Number of Targets. In Proceedings 22nd IFAC World Congress (IFAC-23). 2023, 5629–5634. DOI BibTeX

    @inproceedings{ifac23-bilal,
    	author = "Bilal Yousuf and Zsofia Lendek and Lucian Bu\c{s}oniu",
    	title = "Multi-Agent Exploration-Based Search for an Unknown Number of Targets",
    	booktitle = "Proceedings 22nd {IFAC} World Congress ({IFAC-23})",
    	year = 2023,
    	month = "6--11 July",
    	pages = "5629--5634",
    	address = "Seoul, Korea",
    	owner = "Administrator",
    	doi = "https://doi.org/10.1016/j.ifacol.2023.10.206",
    	timestamp = "2008.03.06"
    }
    
  5. Petar Bevanda, Johannes Kirmayr, Stefan Sosnowski and Sandra Hirche. Learning the Koopman Eigendecomposition: A Diffeomorphic Approach. In 2022 American Control Conference (ACC) (). 2022, 2736-2741. DOI BibTeX

    @inproceedings{9867829,
    	author = "Bevanda, Petar and Kirmayr, Johannes and Sosnowski, Stefan and Hirche, Sandra",
    	booktitle = "2022 American Control Conference (ACC)",
    	title = "Learning the Koopman Eigendecomposition: A Diffeomorphic Approach",
    	year = 2022,
    	volume = "",
    	number = "",
    	pages = "2736-2741",
    	doi = "10.23919/ACC53348.2022.9867829"
    }
    
  6. Matija Sukno and Ivana Palunko. Hand-Crafted Features for Floating Plastic Detection. In IEEE/RSJ International Conference on Intelligent Robots and Systems. 2022. URL BibTeX

    @inproceedings{suknoIros2022,
    	author = "Matija Sukno and Ivana Palunko",
    	title = "Hand-Crafted Features for Floating Plastic Detection",
    	booktitle = "IEEE/RSJ International Conference on Intelligent Robots and Systems",
    	year = 2022,
    	note = "accepted",
    	month = "23--27 October",
    	address = "Kyoto, Japan",
    	url = "https://seaclear-project.eu/images/bibtex/Hand-crafted_IROS_2022.pdf"
    }
    
  7. Ivica Nakić, Domagoj Tolić, Ivana Palunko and Zoran Tomljanović. Numerically Efficient Agents-to-Group H∞ Analysis. In 10th Vienna International Conference on Mathematical Modelling MATHMOD 2022 55(20). 2022, 199-204. URL, DOI BibTeX

    @inproceedings{NAKIC2022199,
    	title = "Numerically Efficient Agents-to-Group H∞ Analysis",
    	volume = 55,
    	number = 20,
    	pages = "199-204",
    	year = 2022,
    	booktitle = "10th Vienna International Conference on Mathematical Modelling MATHMOD 2022",
    	issn = "2405-8963",
    	doi = "https://doi.org/10.1016/j.ifacol.2022.09.095",
    	url = "https://seaclear-project.eu/images/bibtex/Numerically_efficient_agents.pdf",
    	author = "Ivica Nakić and Domagoj Tolić and Ivana Palunko and Zoran Tomljanović",
    	keywords = "multi-agent systems, norm, network robustness/resiliency, output synchronization, linear systems",
    	abstract = "This paper proposes a numerically efficient approach for computing the maximal/minimal impact a subset of agents has on the cooperative system. For instance, if one is able to disturb/bolster several agents so as to maximally disturb/bolster the entire team, which agents to choose and what kind of inputs to apply? We quantify the agents-to-team impacts in terms of H∞ norm whereas output synchronization is taken as the underlying cooperative control scheme. Sufficient conditions on agents’ parameters, synchronization gains and topology are provided such that the associated H∞ norm attains its maximum for constant agents’ disturbances. Linear second-order agent dynamics and weighted undirected topologies are considered. Our analyses also provide directions towards improving graph design and tuning/selecting cooperative control mechanisms. Lastly, numerical examples, some of which include forty thousand agents, are provided."
    }
    
  8. Alexandru Pop and Levente Tamas. Next best view estimation for volumetric information gain. In 6th IFAC Conference on Intelligent Control and Automation Sciences ICONS 2022 55(15). 2022, 160-165. URL, DOI BibTeX

    @inproceedings{POP2022160,
    	title = "Next best view estimation for volumetric information gain",
    	volume = 55,
    	number = 15,
    	pages = "160-165",
    	year = 2022,
    	booktitle = "6th IFAC Conference on Intelligent Control and Automation Sciences ICONS 2022",
    	issn = "2405-8963",
    	doi = "https://doi.org/10.1016/j.ifacol.2022.07.625",
    	url = "https://www.sciencedirect.com/science/article/pii/S2405896322010369",
    	author = "Alexandru Pop and Levente Tamas",
    	keywords = "Trajectory, path planning, perception, sensing, deep neural network",
    	abstract = "In this work we propose a novel next best view (NBV) generation algorithm for volumetric information maximization. The primary data source is a Time-of-Flight (ToF) camera and the output is the next position of the depth sensor that maximizes a chosen score, either coverage or histogram of volumetric estimations for parallelepipedic shapes on the observed scene. Our learning-based method was validated on a large scale of real and synthetic data. The demo code, custom data sets, and videos are available on the author's website."
    }
    
  9. Petar Bevanda, Max Beier, Shahab Heshmati-Alamdari, Stefan Sosnowski and Sandra Hirche. Towards Data-driven LQR with Koopmanizing Flows⋆. In 6th IFAC Conference on Intelligent Control and Automation Sciences ICONS 2022 55(15). 2022, 13-18. URL, DOI BibTeX

    @inproceedings{BEVANDA202213,
    	title = "Towards Data-driven LQR with Koopmanizing Flows⋆",
    	volume = 55,
    	number = 15,
    	pages = "13-18",
    	year = 2022,
    	booktitle = "6th IFAC Conference on Intelligent Control and Automation Sciences ICONS 2022",
    	issn = "2405-8963",
    	doi = "https://doi.org/10.1016/j.ifacol.2022.07.601",
    	url = "https://arxiv.org/pdf/2201.11640.pdf",
    	author = "Petar Bevanda and Max Beier and Shahab Heshmati-Alamdari and Stefan Sosnowski and Sandra Hirche",
    	keywords = "Machine learning, Koopman operators, Learning for control, Representation Learning, Neural networks, Learning Systems",
    	abstract = "We propose a novel framework for learning linear time-invariant (LTI) models for a class of continuous-time non-autonomous nonlinear dynamics based on a representation of Koopman operators. In general, the operator is infinite-dimensional but, crucially, linear. To utilize it for effcient LTI control design, we learn a finite representation of the Koopman operator that is linear in controls while concurrently learning meaningful lifting coordinates. For the latter, we rely on Koopmanizing Flows - a diffeomorphism-based representation of Koopman operators and extend it to systems with linear control entry. With such a learned model, we can replace the nonlinear optimal control problem with quadratic cost to that of a linear quadratic regulator (LQR), facilitating efficacious optimal control for nonlinear systems. The superior control performance of the proposed method is demonstrated on simulation examples."
    }
    
  10. Damir Bulić, Domagoj Tolić and Ivana Palunko. Beam-Based Tether Dynamics and Simulations using Finite Element Model. In 6th IFAC Conference on Intelligent Control and Automation Sciences ICONS 2022 55(15). 2022, 154-159. URL, DOI BibTeX

    @inproceedings{BULIC2022154,
    	title = "Beam-Based Tether Dynamics and Simulations using Finite Element Model",
    	volume = 55,
    	number = 15,
    	pages = "154-159",
    	year = 2022,
    	booktitle = "6th IFAC Conference on Intelligent Control and Automation Sciences ICONS 2022",
    	issn = "2405-8963",
    	doi = "https://doi.org/10.1016/j.ifacol.2022.07.624",
    	url = "https://www.sciencedirect.com/science/article/pii/S2405896322010357",
    	author = "Damir Bulić and Domagoj Tolić and Ivana Palunko",
    	keywords = "tether dynamics, tethered robots, Finite Element Method, beam equation, PDEs, simulations",
    	abstract = "Building on the beam equation, this paper derives a realistic and versatile 3D model of tether/cable/umbilical dynamics and approximates it via the Finite Element Method (FEM). This FEM implementation/approximation can be executed in real time, when the model accuracy is not of prime concern for the application at hand, or in an offline fashion, when the model accuracy is essential. Our model allows for different time-varying forces (e.g., wind gusts, sea currents, waves, etc.) acting along the cable. However, unlike the related works, where tether bending and stretching are mutually independent resulting in unrealistic tether elongations, our implementation takes into account the tether length, which in turn imposes coupling among transversal and longitudinal tether displacements. Our work is motivated by applications of tethered robots such as Unmanned Underwater Vehicles (UUVs), most commonly Remotely Operated Vehicles (ROVs) and Unmanned Aerial Vehicles (UAVs). In such applications, the tether is typically kept under tension leading to “small” tether displacements and the absence of deformations and entanglements, which is why the beam equation is selected to start with. However, partially owing to the attention paid to the tether length, our dynamic model behaves well even for “larger” displacements as illustrated in Gazebo ROS using a tethered ROV whereas the tether itself is implemented in Matlab via FEM."
    }
    
  11. Bilal Yousuf, Zsofia Lendek and Lucian Busoniu. Exploration-Based Search for an Unknown Number of Targets using a UAV. In Proceedings 6th IFAC Conference on Intelligent Control and Automation Sciences (ICONS-22). 2022. URL, DOI BibTeX

    @inproceedings{icons22-bilal,
    	author = "Yousuf, Bilal and Lendek, Zsofia and Busoniu, Lucian",
    	title = "Exploration-Based Search for an Unknown Number of Targets using a UAV",
    	booktitle = "Proceedings 6th IFAC Conference on Intelligent Control and Automation Sciences ({ICONS-22})",
    	year = 2022,
    	note = "accepted",
    	month = "13--15 July",
    	address = "Cluj-Napoca, Romania",
    	url = "http://busoniu.net/files/papers/icons22-bilal.pdf",
    	doi = "https://doi.org/10.1016/j.ifacol.2022.07.614"
    }
    
  12. Jan Brüdigam, Martin Schuck, Alexandre Capone, Stefan Sosnowski and Sandra Hirche. Structure-Preserving Learning Using Gaussian Processes and Variational Integrators. In Proceedings of The 4th Annual Learning for Dynamics and Control Conference 168. 2022, 1150–1162. URL PDF BibTeX

    @inproceedings{pmlr-v168-brudigam22a,
    	title = "Structure-Preserving Learning Using Gaussian Processes and Variational Integrators",
    	author = {Br\"udigam, Jan and Schuck, Martin and Capone, Alexandre and Sosnowski, Stefan and Hirche, Sandra},
    	booktitle = "Proceedings of The 4th Annual Learning for Dynamics and Control Conference",
    	pages = "1150--1162",
    	year = 2022,
    	volume = 168,
    	series = "Proceedings of Machine Learning Research",
    	month = "23--24 Jun",
    	publisher = "PMLR",
    	pdf = "https://proceedings.mlr.press/v168/brudigam22a/brudigam22a.pdf",
    	url = "https://proceedings.mlr.press/v168/brudigam22a.html",
    	abstract = "Gaussian process regression is increasingly applied for learning unknown dynamical systems. In particular, the implicit quantification of the uncertainty of the learned model makes it a promising approach for safety-critical applications. When using Gaussian process regression to learn unknown systems, a commonly considered approach consists of learning the residual dynamics after applying some generic discretization technique, which might however disregard properties of the underlying physical system. Variational integrators are a less common yet promising approach to discretization, as they retain physical properties of the underlying system, such as energy conservation and satisfaction of explicit kinematic constraints. In this work, we present a novel structure-preserving learning-based modelling approach that combines a variational integrator for the nominal dynamics of a mechanical system and learning residual dynamics with Gaussian process regression. We extend our approach to systems with known kinematic constraints and provide formal bounds on the prediction uncertainty. The simulative evaluation of the proposed method shows desirable energy conservation properties in accordance with general theoretical results and demonstrates exact constraint satisfaction for constrained dynamical systems."
    }
    
  13. Tudor Santejudean, Lucian Busoniu, Vineeth Varma and Constantin Morarescu. A simple path-aware optimization method for mobile robots. In 6th IFAC Symposium on Telematics Applications (TA-22). 2022. URL, DOI BibTeX

    @inproceedings{ta22-tudor,
    	author = "Santejudean, Tudor and Busoniu, Lucian and Varma, Vineeth and Morarescu, Constantin",
    	title = "A simple path-aware optimization method for mobile robots",
    	booktitle = "6th IFAC Symposium on Telematics Applications ({TA-22})",
    	year = 2022,
    	note = "accepted",
    	month = "15--17 June",
    	address = "Nancy, France",
    	url = "http://busoniu.net/files/papers/ifacta22-tudor.pdf",
    	doi = "https://doi.org/10.1016/j.ifacol.2022.08.001"
    }
    
  14. Jerome Weston, Domagoj Tolić and Ivana Palunko. Mixed Use of Pontryagin's Principle and the Hamilton-Jacobi-Bellman Equation in Infinite- and Finite-Horizon Constrained Optimal Control. In International Conference on Intelligent Autonomous Systems (IAS). 2022. URL BibTeX

    @inproceedings{jerome2022,
    	author = "Weston, Jerome and Tolić, Domagoj and Palunko, Ivana",
    	title = "Mixed Use of Pontryagin's Principle and the Hamilton-Jacobi-Bellman Equation in Infinite- and Finite-Horizon Constrained Optimal Control",
    	booktitle = "International Conference on Intelligent Autonomous Systems (IAS)",
    	year = 2022,
    	note = "accepted",
    	month = "13-16 June",
    	address = "Zagreb, Croatia",
    	url = "https://seaclear-project.eu/images/bibtex/Mixed_use_of_Pontryagin_ZagrebWTP.pdf"
    }
    
  15. Antun Đuraš, Matija Sukno and Ivana Palunko. Recovering the 3D UUV Position using UAV Imagery in Shallow-Water Environments. In 2022 International Conference on Unmanned Aircraft Systems (ICUAS) (). June 2022, 948-954. DOI BibTeX

    @inproceedings{9836195,
    	author = "Đuraš, Antun and Sukno, Matija and Palunko, Ivana",
    	booktitle = "2022 International Conference on Unmanned Aircraft Systems (ICUAS)",
    	title = "Recovering the 3D UUV Position using UAV Imagery in Shallow-Water Environments",
    	year = 2022,
    	month = "June",
    	volume = "",
    	number = "",
    	pages = "948-954",
    	abstract = "In this paper we propose a novel approach aimed at recovering the 3D position of an UUV from UAV imagery in shallow-water environments. Through combination of UAV and UUV measurements, we show that our method can be utilized as an accurate and cost-effective alternative when compared to acoustic sensing methods, typically required to obtain ground truth information in underwater localization problems. Furthermore, our approach allows for a seamless conversion to geo-referenced coordinates which can be utilized for navigation purposes. To validate our method, we present the results with data collected through a simulation environment and field experiments, demonstrating the ability to successfully recover the UUV position with sub-meter accuracy.",
    	keywords = "",
    	doi = "10.1109/ICUAS54217.2022.9836195",
    	issn = "2575-7296"
    }
    
  16. Marian-Leontin Pop and Levente Tamas. MPI Planar Correction of Pulse Based ToF Cameras. In 2022 IEEE International Conference on Automation, Quality and Testing, Robotics (AQTR) (). 2022, 1-6. DOI BibTeX

    @inproceedings{9802059,
    	author = "Pop, Marian-Leontin and Tamas, Levente",
    	booktitle = "2022 IEEE International Conference on Automation, Quality and Testing, Robotics (AQTR)",
    	title = "MPI Planar Correction of Pulse Based ToF Cameras",
    	year = 2022,
    	volume = "",
    	number = "",
    	pages = "1-6",
    	doi = "10.1109/AQTR55203.2022.9802059"
    }
    
  17. Benjamin Kelényi, Szilárd Molnár and Levente Tamás. 3D Object Recognition using Time of Flight Camera with Embedded GPU on Mobile Robots.. In VISIGRAPP (4: VISAPP). 2022, 849–856. DOI BibTeX

    @inproceedings{kelenyi20223d,
    	title = "3D Object Recognition using Time of Flight Camera with Embedded GPU on Mobile Robots.",
    	author = "Kel{\'e}nyi, Benjamin and Moln{\'a}r, Szil{\'a}rd and Tam{\'a}s, Levente",
    	booktitle = "VISIGRAPP (4: VISAPP)",
    	pages = "849--856",
    	year = 2022,
    	doi = "10.5220/0010972200003124"
    }
    
  18. Vicu-Mihalis Maer, Levente Tamas and Lucian Busoniu. Underwater robot pose estimation using acoustic methods and intermittent position measurements at the surface. In 2022 IEEE International Conference on Automation, Quality and Testing, Robotics (AQTR-22). 2022. URL, DOI BibTeX

    @inproceedings{aqtr22-mihalis,
    	author = "Vicu-Mihalis Maer and Levente Tamas and Lucian Busoniu",
    	title = "Underwater robot pose estimation using acoustic methods and intermittent position measurements at the surface",
    	booktitle = "2022 IEEE International Conference on Automation, Quality and Testing, Robotics ({AQTR-22})",
    	year = 2022,
    	month = "22--24 May",
    	address = "Cluj-Napoca, Romania",
    	url = "http://busoniu.net/files/papers/aqtr22-mihalis.pdf",
    	doi = "https://doi.org/10.1109/AQTR55203.2022.9802002"
    }
    
  19. Vicko Prkačin, Ivana Palunko and Ivan Petrović. Extended Kalman filter for payload state estimation utilizing aircraft inertial sensing. In 2021 Aerial Robotic Systems Physically Interacting with the Environment (AIRPHARO) (). 2021, 1-6. DOI BibTeX

    @inproceedings{9571038,
    	author = "Prkačin, Vicko and Palunko, Ivana and Petrović, Ivan",
    	booktitle = "2021 Aerial Robotic Systems Physically Interacting with the Environment (AIRPHARO)",
    	title = "Extended Kalman filter for payload state estimation utilizing aircraft inertial sensing",
    	year = 2021,
    	volume = "",
    	number = "",
    	pages = "1-6",
    	doi = "10.1109/AIRPHARO52252.2021.9571038"
    }
    
  20. Zewen Yang, Stefan Sosnowski, Qingchen Liu, Junjie Jiao, Armin Lederer and Sandra Hirche. Distributed Learning Consensus Control for Unknown Nonlinear Multi-Agent Systems based on Gaussian Processes. In 2021 60th IEEE Conference on Decision and Control (CDC) (). December 2021, 4406-4411. URL, DOI BibTeX

    @inproceedings{9683522,
    	author = "Yang, Zewen and Sosnowski, Stefan and Liu, Qingchen and Jiao, Junjie and Lederer, Armin and Hirche, Sandra",
    	booktitle = "2021 60th IEEE Conference on Decision and Control (CDC)",
    	title = "Distributed Learning Consensus Control for Unknown Nonlinear Multi-Agent Systems based on Gaussian Processes",
    	year = 2021,
    	volume = "",
    	number = "",
    	pages = "4406-4411",
    	abstract = "In this paper, a distributed learning leader-follower consensus protocol based on Gaussian process regression for a class of nonlinear multi-agent systems with unknown dynamics is designed. We propose a distributed learning approach to predict the residual dynamics for each agent. The stability of the consensus protocol using the data-driven model of the dynamics is shown via Lyapunov analysis. The followers ultimately synchronize to the leader with guaranteed error bounds by applying the proposed control law with a high probability. The effectiveness and the applicability of the developed protocol are demonstrated by simulation examples.",
    	keywords = "",
    	doi = "10.1109/CDC45484.2021.9683522",
    	issn = "2576-2370",
    	month = "Dec",
    	url = "https://mediatum.ub.tum.de/doc/1633342/t79gh8w431q90v5eb50wiwxmf.2103.15929.pdf"
    }
    
  21. Tudor Sântejudean and Lucian Busoniu. Path-aware optimistic optimization for a mobile robot. In 2021 60th IEEE Conference on Decision and Control (CDC). 2021. URL, DOI BibTeX

    @inproceedings{santejudean_Pathaware_14,
    	address = "Austin, TX, USA",
    	title = "Path-aware optimistic optimization for a mobile robot",
    	booktitle = "2021 60th IEEE Conference on Decision and Control (CDC)",
    	isbn = "978-1-66543-659-5",
    	url = "http://busoniu.net/files/papers/cdc21-tudor.pdf",
    	doi = "10.1109/CDC45484.2021.9683546",
    	abstract = "We consider problems in which a mobile robot samples an unknown function defined over its operating space, so as to find a global optimum of this function. The path travelled by the robot matters, since it influences energy and time requirements. We consider a branch-and-bound algorithm called deterministic optimistic optimization, and extend it to the path-aware setting, obtaining path-aware optimistic optimization (OOPA). In this new algorithm, the robot decides how to move next via an optimal control problem that maximizes the long-term impact of the robot trajectory on lowering the upper bound, weighted by bound and function values to focus the search on the optima. An online version of value iteration is used to solve an approximate version of this optimal control problem. OOPA is evaluated in extensive experiments in two dimensions, where it does better than path-unaware and local-optimization baselines.",
    	urldate = "2022-05-05",
    	publisher = "IEEE",
    	author = "Sântejudean, Tudor and Busoniu, Lucian",
    	month = "",
    	year = 2021
    }
    

Read more: Publications

SEACLEAR media kit

Brief description

SEACLEAR — short for SEarch, identificAtion and Collection of marine Litter with Autonomous Robots — is a Horizon 2020 funded project that aims to solve, with the help of robots and artificial intelligence, one of the most important environmental problems: ocean litter.

The problem

Today's oceans contain 26-66 million tons of waste, with approximately 94% located on the seafloor. So far, collection efforts have focused mostly on surface waste. The few efforts to gather underwater waste involve human divers who are put in potential danger.

No solution exists that exploits autonomous robots for underwater litter collection. The SEACLEAR project will develop the first.

The SEACLEAR system

The SEACLEAR project aims at automating the process of searching, identifying, and collecting marine litter, using a team of autonomous robots that work together.

Here is how it works. The waste clearing system consists of underwater robots, a surface vessel and drones working together. First, the underwater robots and the drones use sensor data and artificial intelligence to locate and identify litter on the seafloor. When litter is detected, an underwater robot equipped with a gripper is sent to collect the waste. 

A more detailed description is available here.

About the team

We are a mixed research-industry team with members from: TU Delft, DUNEA, Fraunhofer CML, Hamburg Port Authority, Subsea Tech, TU Cluj-Napoca, TU Munich, and University of Dubrovnik.

The project coordinator is professor Bart De Schutter, from Delft University of Technology.

Funding

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No. 871295.

The SEACLEAR project spans four years, running from January 1st, 2020 to December 31st, 2023.

Quotes

“Coastal waters all over the world are polluted with litter, such as pieces of plastic, bottles or tyres. At the moment, divers are cleaning up this waste from the seabed, especially in tourist areas. However, this is an expensive solution and can sometimes pose dangers for divers. That’s why we’ve joined forces with seven other partners in the SEACLEAR project to develop an autonomous system using underwater robots to remove waste from the seabed. And important to say, we are careful not to affect life on the seabed.” 

Professor Bart De Schutter, SEACLEAR project manager, TU Delft’s Centre for Systems and Control.

“The SEACLEAR system works as follows: We have a  surface vessel on the water and two underwater robots. The somewhat smaller robot is the observation robot. It scans the seabed with a camera and sonar. This robot maps out where the litter is located and what kind of waste it is. The robot can also distinguish between litter and aquatic life, such as fish and seaweed. We use advanced algorithms to make this distinction.

Once the observation robot has identified litter, it sends this information to the other underwater robot, which is equipped with a gripper. This robot goes to the litter and picks it up with the gripper and deposits it into a large basket. The gripper is designed with a frame structure so fish can easily escape when being picked up by accident.”

Professor Bart De Schutter, SEACLEAR project manager, TU Delft’s Centre for Systems and Control.

Photos

Click on a picture to see the full version.

Images and videos may be used freely. Credit: Seaclear Project

Components

  • SeaCAT

  • Tortuga

  • Mini Tortuga in the water

  • DJI Matrice M210

  • Gripper Prototype Test

  • Collection Basket Underwater

  • LARS Mini Tortuga 2

  • Tortuga with the gripper

System

  • SeaClear system

  • Concept Explanation

  • First Litter Grabbed

  • Gripper collecting litter

  • Monitoring Device

  • SEACLEAR Team

  • SEACLEAR system explanation

  • SeaCat with all the components

Videos

 
Please feel free to use our YouTube videos for any material that mentions SeaClear. A couple of representative videos are provided below.
Upon request, we also offer additional and unmarked/raw footage of our real-world trials, robots and system components, etc. Reach out to our press contact — see email below. 

Who talks about us

Press contact

Assistant professor Tassos Natsakis

Technical University of Cluj-Napoca, Romania

This email address is being protected from spambots. You need JavaScript enabled to view it.

Mobile phone: +40753820782

General contact

www.seaclear-project.eu

This email address is being protected from spambots. You need JavaScript enabled to view it.

Linkedin

Twitter

YouTube

Facebook

Read more: SEACLEAR Media Kit

A robotic system intended to clean litter from the seafloor has passed its first real-life tests in waters near Dubrovnik, Croatia. 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 official litter to be picked up from the seabed. This series of tests is a small step in humanity’s goal of cost-effectively cleaning the seafloor, where more than 90% of the sea garbage is found.

picking_first_bottle.png

SeaClear — picking up the first litter from the seafloor

A larger image can be found here

The waste clearing system consists of underwater robots, a surface vessel and drones working together. First, the underwater robots and the drones use sensor data and artificial intelligence to locate and identify litter on the seafloor. When litter is detected, an underwater robot equipped with a gripper is sent to collect the waste. The system has been in development for the past two years by the European-funded SeaClear project (SEarch, identificAtion and Collection of marine Litter with Autonomous Robots) that joins together researchers and industry from five countries.

 This October, three key components of the system were tested in a real water environment: the underwater inspection robot, a gripper prototype, and an aerial drone. “We went to a tourist area in Croatia, Lokrum Island near Dubrovnik, which is one of the first places we want to clean once the SeaClear system is fully operational”, says professor Bart De Schutter, from Delft University of Technology, the Netherlands, who leads the project. “The main star of the test was the underwater inspection robot. We wanted to see whether it’s able to detect and move towards the litter in various water conditions”, explains De Schutter.

 mini-tortuga-underwater.png

SeaClear — Underwater inspection robot

A larger image can be found here

 The first live test resulted in the first official litter being picked up from the water by the SeaClear robot: a plastic bottle! This happened during the test at a second location, also close to Dubrovnik, Croatia, but with less clear water. The team tested a 3D printed prototype of its newly designed gripper, to identify weak spots in the design and operation under real-life conditions and indeed, it broke. However, to have a proof-of-concept, the researchers were able to repurpose an existing, sturdier gripper, and managed to successfully pick up litter underwater.

 The test was also important for the software part of the project. “The concept is simple on paper and not unlike our home robot vacuum cleaners”, says professor Lucian Bușoniu, from Technical University Cluj, Romania, who is leading the litter-search part of the project. “However, the complexity of robot positioning and control, as well as of litter identification and collection, is much higher and makes the objectives much more difficult to achieve. We wanted to test out our concept and we now need to go through all the data sets from our cameras, the sonar and the robot positioning sensors. This will help improve our software.”

 A completely autonomous robotic system for sea litter cleaning would be a vital and cheap solution to one of the largest environmental problems our planet is facing. Today's oceans contain 26-66 million tons of waste, such as pieces of plastic, bottles or tyres. The waste that we see at the surface represents a mere 6% of the total, the rest is resting at the bottom. Tourist resorts and port authorities currently resort to divers to clean up this waste from the seabed, sometimes at the cost of their safety.

 “For example, in September 2020, a human-only marine litter clean-up that took place in the same test area close to Lokrum Island, resulted in 17 divers collecting more than 88 kg of waste from the seafloor in a one-hour period”, said Iva Pozniak, principal investigator for Regional Agency Dunea, Croatia. When completed, the SeaClear system should achieve similar results with minimal human intervention.

 

SeaClear is a European Horizon 2020 project that was launched on 1 January 2020. It runs until December 2023. The total budget is approximately EUR 5 million. There are eight partners from five countries and 49 researchers involved.

The eight partners are: TU Delft, Hamburg Port Authority, TU Cluj-Napoca, Subsea Tech, TU Munich, Fraunhofer CML, University of Dubrovnik and DUNEA.

Further information

A more detailed and technical description of how the system works is available here.

 

Images:

https://drive.google.com/drive/folders/1buVA-u3_OVMi5kOLSaq52boT_jnqYM9g

 

Video:

Tags: press release

Read more: First successful tests of robotic system for...

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No. 871295.

The SEACLEAR project spans four years, running from January 1st, 2020 to December 31st, 2023.

SEACLEAR NEWSLETTER

Stay up to date with news about the project. Expected frequency for the newsletter is twice a year.

SEACLEAR is not responsible for the content of third-party sites to which we link from this site.


Design & maintenance by Lichtbringer