The ship you can see in the picture has cruised around the fjord outside Trondheim for the past few summers, usually closely followed by students from NTNU. Despite being only three metres long and just over half a metre wide, ReVolt has attracted the attention of many large players in the shipping industry.
"The reason for us developing ReVolt and focusing such a lot on autonomy is that we want to learn. Not because we have plans to compete with the shipping industry in this field," says Tom Arne Pedersen, a principal researcher with DNV GL. "Our main purpose with ReVolt is to ensure that the autonomous systems released into the market in a few years are as safe as or safer than conventional ships," he says.
Autonomous container shipsLet us go a few years back in time to when, with the growth of artificial intelligence and machine learning, the idea of autonomous vessels, either on land or at sea, left the field of science fiction and become a real possibility. DNV GL – a company that has carried out the classification, testing and verification of ships for 155 years – realised it had to be at the forefront. "We quite simply had to know how these autonomous ships would function, the requirements we had to stipulate and how we were to test these ships," says Pedersen.
Reason for the development of the ReVolt conceptThis started in 2014 as a small container-ship concept – 60 metres long and with a maximum speed of six knots, environmentally friendly and energy-efficient, for use in coastal areas. The concept included autonomous steering technology, i.e. the ship was to be self-sailing and recharged every time it was in port. The concept was never meant to be built, but a scale model was constructed and sails around the Trondheim harbour basin four years later.
Exploring the technology
This model project gives DNV GL an opportunity to demonstrate autonomy, explore the opportunities afforded by this technology and develop a test platform for these systems. The company does so in close collaboration with NTNU, which has had master's degree students from the marine technology, robotics and cybernetics departments in summer jobs with Pedersen for several years. These students have also written project and master's theses in cooperation with DNV GL.
An important part of the ReVolt concept is the ship's digital twin. This is a full model of the ship and the systems on board and is run on both an office computer and DNV GL's cloud solution. It is used by the students to try out their solutions before testing them on the ship in the water. "We also use this to develop our own test tools," says Pedersen. The work on the test tools has revealed some of the challenges that will be faced in the future.
"Sea rules are written for human operations, with room for interpretation, while a computer needs completely specific rules. Here, machine-readable rules must be developed and that is something that industry and the shipping industry – those who are actually going to build and operate the ships – must do together," advises Pedersen.
Autonomous ships within 10 years
Pedersen predicts that autonomous ships may become part of the logistics network along the coast within 10 years. "I don't think all ship traffic will be replaced, but we can see some areas that will probably change: such as the coastal transport of goods. The major potential here lies in autonomous freighters replacing a lot of the road haulage on land, as Yara and Kongsberg are doing with the Yara Birkeland. Although the Yara Birkeland will be delivered as early as in 2020, this is as a pilot project, and it will probably take a few years before a large number of autonomous ships can take part in a large logistics network," says Pedersen. He believes the biggest remaining challenge is the ships' 'ability' to see and understand the world around it, also called situation understanding.
"Just imagine when there is thick fog and rain, everything is much more difficult. Currently, a lot of manual adjustments are done to make sensors function optimally. In addition, as a human being you can use your hearing and experience in a different way to at least the autonomous systems we currently have. "Another challenge is how to make the autonomous systems operate so that we humans can understand and trust their decisions.
"However quickly developments take place, there will be a long period when we have conventional and autonomous ships in the water at the same time. How can we make sure they can cooperate and trust each other?" asks Pedersen. Will the captain disappear? "Like many other technological shifts, I think this will lead to some types of jobs disappearing, while others will arise. We will still have captains, but maybe they will go to work in a control room onshore and be responsible for several vessels."