With talk of driverless cars and remote drone aircraft all the rage, it’s little wonder that autonomous ships have become the ‘hot topic’ of maritime research and development.

The idea of unmanned ships might have seemed fanciful a couple of years ago, but since then snowballing interest in the possibilities of ‘smart’ or intelligent shipping have laid the groundwork for just such an eventuality.

At the same time, projects to develop autonomous prototypes and the new technologies they will require have proliferated at a rate that few would have predicted. Countries like Norway, Finland, the UK, Singapore, Japan and China have all set up broad-based R&D programmes and are vying for leadership in the field, with model testing and in some cases sea trials of technology demonstrators already underway.

Different levels

Some definitions first: the automotive and aviation industries have laid down five different stages of autonomous operations, that the shipping industry has broadly adopted, which in ascending order are:

Level 1 – a ship that can benefit from a remote operator’s assistance (Driver Assistance);

Level 2 – a ship capable of being partly or periodically left unattended (Partial Automation);

Level 3 – a ship with an automated drive system that can self-drive providing an operator can step in as required (Conditional Automation);

Level 4 – as with the previous level, but capable of self-driving if an operator does not step in (High Automation); and

Level 5 – a ship that can self-drive totally unmanned in the same conditions and with the same capability as if it were manned (Full Automation).

Objectives

Full autonomy is the ‘holy grail’ of this technological quest, and many feel it could help eliminate the approximate 80% of maritime accidents attributed to human error as well as contribute to greater efficiency.

But others argue that widespread adoption of Level 5 may not be practicable or societally acceptable in the foreseeable future – except perhaps in a few, highly specialised operations such as ferry crossings of remote fjords or hazardous duties such as fire-fighting.

However, even sceptics concede that the journey towards autonomous shipping – just like the ‘space race’ of the 1950s/60s before it – may produce technological spin-offs of as much benefit as the destination itself. It can also lead to the de-manning of certain ship operations, and hence cost reductions for operators.

With technological innovation advancing at such an exponential rate, no one is going to rule out the possibility of unmanned ocean-going ships altogether.

Research projects

The technical background to research into autonomous ships is that “advances in sensor technology, data analytics and bandwidth to shore are fundamentally changing the way shipping works,” says Dr Pierre C. Sames, director of Group Technology & Research for classification society DNV GL. “And as operations are digitalised, they become more automated.” Progressive levels of automation in turn lead on to autonomy.

DNV GL is involved in a whole range of projects on ship autonomy including the Norwegian Forum for Autonomous Ships (NFAS), which is enthusiastically supported by the Norwegian Government. Norway has even earmarked a stretch of water off Trondheim for the testing of unmanned vessels.

Then there’s the Advanced Autonomous Waterborne Applications Initiative (AAWA), another Scandinavian-led effort looking into a whole range of issues around autonomous shipping including safety, legal, economic and societal acceptance aspects, as well as the actual technologies required.

DNV GL is also certifying an electric container ship that automation specialist Kongsberg and chemicals group Yara plan to build by the first quarter of 2019, as well as model testing its own concept for a battery-powered, fully autonomous feeder boxship for operation in Norwegian fjords, dubbed ReVolt. It should be noted that electric ships avoid the risk of bunker oil spills in the event of a collision or grounding, which can be seen as a major obstacle to unmanned operations by conventionally fuelled ships.

In addition, the class society is involved in several joint R&D projects to develop the ancillary technology required for remote shipping, such as Autosea – on sensor fusion and collision avoidance – and ROMAS (Remote Operations of Machinery and Automation Systems) – for on-shore engine control room technology.

Finnish-based companies are meanwhile heading up a research project called Ecosystem for Autonomous Ships, a collaborative platform that aims to introduce the world’s first autonomous marine transport system by 2025.

Then there’s a UK-Norwegian collaborative project called SIMAROS (Safe Implementation of Autonomous and Remote Operation of Ships), which plans to build a fully autonomous offshore vessel, Hrönn, at Fjellstrand shipyard in 2018.

Another company leading the charge towards autonomous shipping is Rolls-Royce (R-R), a pioneer of autonomous naval craft which tug operator Svitzer used to carry out what was hailed as the world’s first remotely operated commercial vessel manoeuvre in Copenhagen harbour earlier this year.

Working from a Remote Operating Centre (ROC) in Svitzer’s nearby HQ, the master of the 28-metre tug Svitzer Mermod – fitted with a R-R Dynamic Positioning System connected to a remotely controlled system – undocked the vessel, turned her through 360 degrees, piloted towards the Svitzer HQ and then docked again. Sensors aboard the vessel provided the ROC with situational awareness throughout.

Several months of further testing of remote and autonomous operations aboard the vessel are now taking place, with technical assistance from UK class society Lloyd’s Register (LR).

The Marine division of R-R says it is also looking at opportunities to develop its ship intelligence technology in Singapore, where it has a strategic partnership with Nanyang Technological University (NTU).

Separately, the Maritime and Port Authority of Singapore (MPA) has signed a Memorandum of Understanding (MoU) with technology firm Alpha Ori, LR and the National University of Singapore’s Technology Centre for Offshore and Marine (TCOMS) to look into opening a Centre of Innovation for joint research in robotics and unmanned systems in port and shipping for operations, maintenance and inspections in hazardous zones, as well as maritime cybersecurity.

Elsewhere, Japanese shipping company NYK (Nippon Yusen Kabushiki Kaisha) says it plans to test an autonomously operating containership with the crew on standby in the Pacific Ocean in 2019, while an Unmanned Cargo Ship Development Alliance has been set up in China – also including overseas partners ABS, R-R and Wärtsilä – which plans to deliver an unmanned cargo ship by October 2021.

Technology aside, autonomous shipping still faces regulatory, legal and safety and security issues that remain to be resolved.

On the regulation front, the UK Maritime Autonomous Systems Regulatory Working Group (MASWRG) has developed a Code of Conduct for Maritime Autonomous Surface Ships (MASS), which it submitted to the 98th session of the International Maritime Organization’s Maritime Safety Committee in June.

As a result, the IMO has agreed to look into a whole range of issues surrounding MASS, including the human element, safety, security, interactions with ports, pilotage, responses to incidents and protection of the marine environment. The IMO notes that this work could result in amending regulations which, as currently drafted, preclude autonomous or unmanned operations.

The Maritime Safety Committee also agreed that proper consideration should be given to legal aspects, including where the responsibility would lie in case of an accident involving a MASS, its consequences to the cargo, and also the implications to the shore-side.

Finally, perhaps the biggest stumbling block to autonomous shipping might be the cybersecurity risk of a fully automated ship shutting down as a result of a technical malfunction, involuntary or otherwise.

In short, there is plenty still to be discussed about autonomous shipping, with those talks set to involve interested parties throughout the shipping industry for many years to come.

“The simulation of human intelligence processes by machines, especially computer systems” is how Artificial Intelligence (AI) is defined by classification society Lloyd’s Register in its latest report on Autonomous Systems, co-authored by QinetiQ and University of Southampton, published this September.

AI technology already forms part of our daily lives, points out the report, in the virtual assistants on smartphones like Siri, next purchase prediction by e-retailers like Amazon, and ‘smart’ homes where a computerised control system can learn your behaviour patterns and adjust heating and lighting systems accordingly.

For autonomous shipping, however, AI will have to develop further if it is to become a vital enabling technology, capable of providing instant reactions to unique and potentially catastrophic situations.

On manned vessels, it will be able to support bridge command decisions in an advisory role by passing live, contextualised information to the crew as required – a bit like a smartphone virtual assistant

On unmanned vessels, AI will have to go a step further and “use compiled information passed through machine-learning algorithms in order to make a decision, and then act upon it in a timely and correct fashion, enabling [fully] autonomous operations,” says LR.

Early examples of the use of machine learning and algorithms were computers programmed to beat grandmasters at games like chess or Go. By analysing countless thousands of past games they were able to predict behaviour patterns and choose the best move to make – a process known as reinforcement learning – thereby outwitting human intelligence by sheer dint of number-crunching capacity.

But for fully autonomous shipping, AI will need to juggle a far greater number of variants and be capable of object recognition and situational awareness, assessing factors such as other seagoing vessels and their trajectories, navigation channels and rules, sea state and weather conditions, any particular hazards such as reefs, and so on.

As someone once remarked: ‘navigating a ship isn’t rocket science… it’s much more difficult than that.’