AI in port and maritime research in Leiden, Delft and Rotterdam

From a ship that has been designed to tell you what maintenance it needs and when, to an intelligent journey planner for global goods transport. The three universities in Zuid-Holland are abuzz with AI research in the field of ports and maritime. Three researchers explain. Part two in a series of five about themes into which the three universities conduct AI-related research.

‘The big challenges facing ports are accessibility, sustainability and finding the right employees,’ says Rudy Negenborn, Professor of Multi-Machine Operations & Logistics in Delft. ‘In a busy port, you have to optimise your planning to avoid delays, congestion and unnecessary emissions. This doesn’t just require solutions to technical challenges: a solution can only be implemented if an organisation wants and has the right infrastructure for this.'

'We can develop technology that makes it possible for ships to enter a port safely and autonomously, so an operator only has to intervene in an emergency. But what would this mean for the operator’s stress levels? How many ships can they monitor at any one time? And would the ships interact and exchange information with one another?’

All expertise needed

Negenborn has spent about a year and a half intensifying the collaboration with the universities of Leiden and Rotterdam. He can see how the expertise of all three universities is needed when working with many different parties on solutions relating to smart maintenance, smart transportation and smart logistics. ‘In black and white terms, Delft’s main strength is technology, Rotterdam’s is economics and business models, and Leiden’s is social and legal aspects.’

Black and white indeed, because Thomas Bäck from Leiden doesn’t fit this stereotype. He is the head of the Natural Computing research group at the Leiden Institute of Advanced Computer Science (LIACS), which is working on algorithms that apply principles from evolution and nature. A key question in his work is how an algorithm can weigh up different aspects, for instance costs and sustainability, before suggesting the best option. ‘This could mean the best way to design ships, but learning algorithms can also predict the best time for maintenance. This is something that Delft is also researching.’

How we can make a difference

Bäck gained massive inspiration from a focus group with about ten participants from the three universities that focused on AI research in the area of ports and maritime. ‘We discussed how we could really make a difference, while taking the UN Sustainable Development Goals into consideration, for instance. We have to think about social aspects: regulation of AI and how it can explain itself better so as to win people’s trust. The economics and logistics perspectives from Rotterdam were also new to me. It was great.’

More collaboration in Zuid-Holland will create opportunities, says Rob Zuidwijk from Erasmus University Rotterdam. As Professor of Global Supply Chains and Ports, he can see chances to improve the efficiency of goods transport and ports. ‘I want to bring together the public, the city, researchers and not-for-profit organisations in society.’

What autonomous shipping means

If it were up to Zuidwijk, he would invite all of the above to the campus to simulate the Port of Rotterdam in a digital twin environment. ‘Then they would see what autonomous shipping means and what impact certain behaviours have on the employees and public. This would strengthen the links between public and private stakeholders, both at the port and in the city. The dialogues with, among others, Rudy Negenborn and Mark van Koningsveld from Delft University of Technology are helping enormously to get a clear idea of what needs a digital twin environment must meet to get multidisciplinary research with an impact off the ground.’

Collaboration with business is hugely important if you are to have an impact as a researcher, and this is something all the researchers have plenty of experience of. With the Port of Rotterdam of course, and Bäck also works with companies such as Air France-KLM, Tata Steel and the Honda Research Institute in Germany. Negenborn: ‘If we three universities can scale up, it will be easier to reach the boardroom of big companies, which will mean bigger budgets to work with. This will make us more decisive.’

Below is an example of the work of each of the three professors: AI-related research in ports and maritime.

Proactive ship maintenance: algorithm prevents damage and costs

Delft University of Technology - Rudy Negenborn

‘You buy a ship for 30 years. In that time things break, at unfortunate moments. Being at anchor can cost tens of thousands of euros per day, and if a part isn’t in stock... Together with Leiden we are working on health monitoring systems in the Autonomous Shipping research lab (RAS). These kinds of system closely follow how a ship responds to the rudder, the engine speed and so on.
‘The algorithm collects and interprets this data. A sudden reduction in response to the rudder can be caused by sticks around the propeller. If the system suggests this diagnosis, a zigzag manoeuvre can solve the problem. The system can automatically decide whether to temporarily reduce speed, for instance, to prevent engine failure.

‘In this project we in Delft are working on ship dynamics: how to design a ship. Leiden is dealing with the data-science side: how and which data can lead to a diagnosis and advice from the system. A user-friendly dashboard is also coming from Leiden. We’re working on the algorithms together.’

An intelligent journey planner for goods transport

Erasmus University Rotterdam - Rob Zuidwijk

‘Fresh cargo shouldn’t be kept waiting because the cooling system is running. If lorries drive in convoy, they can save as much as 10 percent on fuel. If they coordinate their trips, they can save even more. Inland shipping could be much more economical, but the best use isn’t always made of it. It requires more logistical coordination than freight transport by truck. What we need is a journey planner for European goods transport.

‘Where the 9292 [a journey planner for public transport] brings together all the individual timetables, our system has to be smart. It has to harmonise an enormous amount of data, recognise patterns and then suggest and organise the cheapest, cleanest and fastest route. If something goes wrong, like the recent Suez Canal blockage, the system has to respond.

‘With econometrics and computer science professor Rommert Dekker and professor of goods transport and logistics Lorant Tavasszy from Delft, I’m now thinking about such systems. We’re in talks with parties from the Port of Rotterdam. If the port provides data that show what is happening in the port and in the European hinterland, we can get an algorithm to learn from this. In future all the parties in the future will be able to organise the optimal transport on a virtual map and determine whether the contents of container X will arrive fresh.’

Algorithm shows: the best design and clean your ship in time

Leiden University - Roy de Winter, PhD candidate supervised by Thomas Bäck

‘I’m researching efficient optimisation algorithms at C-Job Naval Architects and Leiden University. An important question is how algorithms can weigh up the lowest production and usage costs against the best quality and result. If, for instance, you make a ship long and slim, it will be more efficient, but you need more steel to build it than to build a shorter, wider one. And more sustainable materials and cleaner fuels are becoming increasingly important in the ship design process.

‘We are also analysing operational data from ships. We recently made a unique discovery. A client wanted to see the influence of the wind on the horsepower required and provided datasets for us to find out. We checked the data from two identical ships and saw that one was over 30 percent more economical than the other. The first one had just been cleaned, whereas the second one had a significant coating of algae and barnacles. That proved to be a significant cause of the difference. Cleaning can be a whole operation, but it’s definitely worthwhile! Nowadays there are methods to do this cleaning in water, with a kind of robot lawn mower, for example.’

Five themes packed with AI research in Zuid-Holland

This article is part two in a series in which we show how teaching and research using or into AI plays a role at Erasmus University Rotterdam, Leiden University and Delft University of Technology. The articles will cover the five themes that the universities are working on, together and alone:

Text: Rianne Lindhout