Defining autonomy – or not

Asked to provide his definition of autonomy, Pierre Sames, Group Technology and Research Director at DNV GL, one of the global bastions of technical knowledge, politely declines.

“DNV GL does not have a definition of autonomy levels,” he says. “Why limit ourselves to such a definition so early in the game? Who knows what it will become?”

  • Pierre Sames, Group Technology and Research Director at DNV GL

He tells that DNV GL, a globally leading quality assurance and risk management company, tends to look at projects first and consider elements of autonomy as needed. “A particular case may feature different combinations of technologies, from fully autonomous in one area to remote control in another. We have to look at the outcome, and a strict definition of autonomy levels would limit us.”

With a PhD from Hamburg University, Sames was many years with the classification society Germanischer Lloyd, and, after the company merged with DNV, continued in DNV GL. His focus throughout has been on engineering and research management.

“Rather than defining autonomy levels, we have defined ship functions,” he says. “Navigation and power or propulsion are the most likely areas of attention now. For each you can pick different approaches: automation, remote, or autonomy.” For example, he cites that conventional navigation can be combined with a fully automated engine room. “So why do we need a definition?” Besides, he jokes, “The truly autonomous ship would probably just set sail for the Caribbean.”

Planning ahead, acting now

Like most in the industry, Sames believes the reality of ships making decisions without humans, at least larger ships, is a long way off. “For navigation, most ships will be on a leash, just not all the time. The ship must do what is required to remain safe, or follow pre-programmed instructions. The main requirement is to achieve minimum risk states.” Short sea and shuttle ships are the more likely candidates in the near future, he says, operating in restricted areas.

“For now, our aim is to help our customers sort out their ideas, to arrive at a solid concept of operations. This concept has to detail the operational scenario: where are they sailing, what is their purpose, how do they operate? What is the safety philosophy and which maintenance is required?”

Make the best of what you have

The main advice from Sames is not to take too large steps. “Use what you have. Regardless, you can only apply what has been through the process, from flag, to class, to owners and OEMs. It is essential to have processes in place to guide the progress.”

Though other industries are progressing on autonomy, Sames sees no real transfer of experience taking place between the automotive and aviation industries and marine. “Navigation systems and situational awareness share some of the same technology, so we could transfer basic algorithms, but they would need to be retrained for a maritime application. And sensor packages would likely be different as well.”

Machine vs. maker

One of the long-standing questions in the autonomy debate is what happens when humans meet machines. “We know how our counterparts will react, but not how a machine would. Reaction times and assumptions are not the same. It is impossible for a human to anticipate how a machine will react, and this is a major barrier to implementing autonomy in a safe way.”

He refers to route exchange in shipping as one way to reduce uncertainty when two vessels interact. “The purpose is to inform surrounding vessels of your intentions. Two merchant ships will follow normal rules of engagement. When a vessel approaches from starboard, you need to go astern. We would need to ensure that autonomous ships follow the same rules as long they navigate in the same waters as conventional ships.”

In such cases it would be better to have a system to guide the interaction also on conventional ships, he argues. “Combined with cameras and other sensors, this would provide a safety boost. Decision support and route exchange can then also improve the safety of the current fleet, not just in future autonomous ships.”

The value of autonomy

The relative value of autonomy is also different for different ships, Sames points out. “It would have a higher impact on smaller, slower ships than on larger, faster vessels. And let’s not forget that there are other ways to save cost, like digitizing paperwork and moving administrative tasks to shore.

“Operators are having problems finding people to work on ships. A shore-based work environment is widely seen as more attractive, so highly automated and remotely operated ships could provide new and attractive shore-based jobs for highly qualified employees.”

System engineering and software engineering also pose challenges to progress toward autonomy. “Capability among players is a major issue. The OEMs are better situated than most yards to take the lead here, because OEMs understand system integration, with capabilities in both systems and software engineering.”

The road unfolds before you

So how does Sames believe the shipping industry will react to the changes that growing autonomy is bound to bring? “First the yards will be affected. Then the most traditional operators, those without the competence or the capacity to operate autonomously. New business models will also emerge, with power and potentially ships as a service, and many more ‘pay-for-play’ solutions.” Reiterating his long-term perspective, Sames concludes: “It will not impact all segments at once, and some perhaps not at all.”

He cites an example from the past: “When LNG as a fuel was being explored, in 2001 Norway started using LNG for ferries. No international regulations or class rules existed at the time, so they started working to develop early class guidelines. The IMO processes started in parallel to this. Six years later, the IMO had interim guidelines, and another 10 years later the IGF code entered into force, setting out internationally accepted requirements.”

Sames believe it is reasonable to expect a similar timeframe for autonomy. “Autonomy is more complicated, but more players stand to benefit.” These factors could balance each other out in determining the time required for industry uptake, he proposes. “But if we are going to find a way forward, we need to have pilots on the water. We need to keep taking small steps, each providing value in themselves, to find out where we want to go with this opportunity.”


Contact us


Share this article

Facebook LinkedIn Twitter WhatsApp