Autonomous ships: When technology is not enough

A crewless, unmanned cargo ship carrying thousands of containers, going anywhere you want, requiring only one manager in a remote location. Sounds pretty nice, right? Lower shipping fees, automated and digitalized voyage planning resulting in higher security on the seas, and crew members staying with their families at new year’s eve. The benefits for commercial shipping are there. In terms of security, autonomous vessels can be a curse and a blessing. No longer can crew be taken hostage by pirates causing a critical situation for the shipowner and its government, but on the other hand, who can defend the vessels without someone on board? Accompanying autonomous vessels, perhaps, but will these be able to distinguish attackers from peaceful nearby ships?

Autonomous or semi-autonomous ships have many possibilities, and the topic make up for a big part in the discussion both inside and outside the shipping industry.

Last year Rolls-Royce said that they were very far in the process of fully autonomous ships, claiming that it could be a reality in 2020 with several unmanned ships at sea.

Earlier the US Navy had demonstrated its interest in the crewless business by unveiling its own self-sailing warship that could track enemy submarines. The ship was part of the bigger strategy with increased usage of drones in its operations. Comparing to the overall popularity of drones on land and at sea, it makes only sense that drone-ships are becoming a reality.

In Amsterdam self-sailing boats might increase the tourist experience of the city with the canals. And while Rolls-Royce is aiming at taking over commercial shipping in the long run, currently smaller vessels are already experiencing practical implementation at sea.

A highly technological remote control center is being developed in Norway , and remote engine and equipment monitoring will enable engineers to supervise remotely. 360 degrees cameras will give the remote operators the required vision and situational awareness.

Rolls Royce control center from their promotion video

The technology for the autonomous vessel

Many will claim that the technology for the autonomous vessel already is available. Currently cargo vessels in fact utilize semi-automating equipment for decreased crew engagement. Autopilots and increased connectivity between all units on board create an environment where data assists the voyage. Technology has decreased the quantitative requirement of crew on board for decades, even more so with the recent years’ technological developments. One technical aspect of autonomous ships is the avoidance of collision with other ships. This in a way correlates with self-driving cars not deviating from its lane and not crashing into other cars.

What is interesting is the fact that technology for transmitting ships’ data such as speed and direction to other ships is already existent. Today both marine radars as well as AIS (Automatic Identification System) are used to read nearby vessels’ voyage along with their data such as owner and where the vessel is registered (its flag). The AIS are mandatory for vessels above 3000 gross tonnage and is today installed on more than 400.000 ships. Data is transmitted via a Very High Frequency (VHF) radio system, and is thought as both a support to traditional radars, but also a tool for situational awareness when at sea. Sea farers utilize this information for avoiding collision, and this technology can be considered the start of autonomous vessels. According to the IMO, AIS-equipped vessels are required to both send and receive their data with other vessels as well as with shore based facilities .

An example of how AIS works

The situational awareness is a key area when thinking autonomous ships. The equipment for this purpose is already a big part of maritime navigation which is a big advantage compared to other industries, for example the car industry, where the currently used navigation tool is a simple handheld GPS.

But the maritime technology still has a long way to go. In 2013 researchers found serious security holes in the AIS both in terms of hardware and software . With intruders being able to manipulate navigational equipment to the extent that Trend Micro did, it shows the dangers of depending on technology. For example the team managed to remove ships from the AIS of other ships, meaning that they were not visible through the AIS. Furthermore the researchers managed to spoof and overload ships’ AIS remotely using basic equipment for a price below 1.000 euros. Furthermore could the AIS be programmed remotely with false information thus transmitting false data to other neaby vessels.

A demonstration of the AIS hack

It has been known for many years that pirates raise money for their operations by including investors . Pirates raising money for a hacker or the education of a hacker to make ships disappear on the radar is not an impossible task – especially if shipowners in a world of fully autonomous ships depend on the ability to see nearby ships to avoid collision. Piracy is big business , but it is not the starved young men with AK 47’s that get rich from the activities. The criminal operators take from 30-75% of the cut, making it a business with a very high margin, thus inciting the search for new business opportunities for example through digital means. It is not only technology that could limit a fully-autonomous shipping industry. In the car industry, the sociological, political, and ethical challenges has been widely discussed, and the topics have relevancy for the maritime industry as well.

For the successful development of the self-driving car, Machine Learning and Artificial Intelligence needs to be, and is very much, successfully a part of the automotive industry. Technology has seen a significant development in software in the latest years. Technologies are becoming incredibly advanced through data research. Google, Facebook, Amazon are investing heavily in data research, and automobile manufacturers are now looking to utilize this in their industry. With applied data research utilizing AI and Machine Learning, the possibilities for the future of cars seem endless. No more delays for the airport, no more drunk drivers, no more careless people speeding in the middle of the city. High Definition maps, high sensitivity cameras and millions of lines of algorithmic programming is not innovative on its own, the technology is there, or we are very close to having it.

A self driving car at Google headquarters

But the successful implementation of autonomous cars requires a complete disruptive mindset of societal structures.

As argumented by Darell M. West, each region and country faces its own problems for the future of autonomous cars. These can be political, regulative, structural, sociological etc. Taking this idea further, the perspectives of autonomous cars will present a series of demands for the governmental structure of any country. Maintenance work on highways will require synchronization with all cars that need to pass through, in order to make sure that speed limits and closed lanes are updated in the cars’ database. Furthermore if the situation where all cars become autonomous, traffic policy will have undergone serious disruption, and the norm will be the autonomous car. Autonomous vehicles is not only a matter of having the technology available, but rather a new societal norm is required for a successful adaption.

The challenges of the adaption to the autonomous automotive vehicle has also been viewed through ethical responsibility. Both pre-programming worst case scenarios and having the vehicle learn from fatal mistakes might present the programmer with unanswerable questions or will be solved on the expense of human lives. Several ethical questions are raised by Patrick Lin in terms of ethical challenges for autonomous cars . If a car for example can foresee a crash a few seconds before it happens but cannot stop, it might be presented with a choice of action. For instance: If the car continues, it will kill and elderly woman and a child. The car could however decide to hit either of the elderly woman or the child thus saving the other. Hitting either raises a series of ethical criticism. First of all programming the car to try and safe a pedestrian by hitting another will not require a program of avoidance but a program of targeting. The car is purposefully killing a person. Next, of course there is the question of who should be saved, the child or the elderly woman. But the ethical questions cannot be answered. Let us say that the car could save 1000 elderly women by killing one child, or the car could save 1000 children by killing 10 previously convicted citizens. The questions need discriminative answers, and this might not create an overall improvement of society with autonomous cars.

The self driving car will have to make some disgusting choices.

Several of these questions can be considered in the shipping industry. First of all the shipping industry is very much concerned about safety and ecology despite horror stories with Maersk irresponsibly recycling its ships in India, BPs oil spill, and the trivial fact that the 10 biggest container ships in the world pollute as much as all the cars in the world. However, the shipping industry is working together under International Maritime Organization (IMO) to make sure the industry as a whole acts responsibly environmentally and safety-wise. Container ships will be detained, if their emergency-distress equipment is not operational. For the shipowners it could mean hundreds of thousands of dollars, if the crews safety is compromised. Furthermore the IMO has worked towards the 2017 implementation of new requirements of Ballast Water Treatment. This will ensure that marine species are not carried through the ballast water to other regions where they might cause unbalance in the ecosystem.

And speaking about safety - both human and environmental, the shipping industry faces its own challenges for unmanned seagoing vessels. In the sciences they speak about the human error, but let us not forget mechanical or digital errors. Often the human error is accounted for beforehand: If you are colorblind, you are not allowed to fly an airplane, when you reach a certain age you are not allowed to drive a car, if you get stress you might be ordered a vacation by your doctor etc. But the mechanical and digital error cannot be accounted for at the same level. First of all in terms security, you cannot safeguard yourself from all possible errors or faults. This is neither possible for people, but we are able to react and make a decision based on similar experiences. The unforeseen must be accounted for in some way for unmanned ships. For the shipping industry it requires not only Machine Learning but AI on a very high level.

Secondly also related to security comes naturally cyber security which has also been discussed in relation to autonomous vehicles. Autonomous systems, cars and ships, require cyber security that can protect the systems from intruders. In terms of cyber security, everyone makes mistakes, and no one is completely safe (not even a presidency candidate). As it looks as of now, self-driving cars can be overtaken by remote attackers, and with this reality, the results could be devastating. The terror acts in Nice, France and Berlin, Germany illustrate that cars can become weapons of terrorism. These were tragic events, and if we consider a hijacked ship where remote attackers overtake the ship and where there is no one onboard to turn off the digital framework and steer the ship manually, the consequences could be fatal. Extortion could cause shipowners to pay millions of dollars rather than having the intruders steer the ship into a disaster. The consequences would be critical. Furthermore, as exemplified with 9/11, hijacking transportation vehicles of any sorts can cause harm and chaos.

Hybrid warfare

Hybrid Warfare has become a buzzword for the new way of war. This new way of war has flowered from the possibilities of the digital age. Russia has been accused of conducting cyber warfare in both Georgia, and Estonia, and its interference in the American president election has also been considered to be hybrid war, and America has been considered to also be in the cyber warfare game. The weaponization of non-weapons is a big part of hybrid warfare. The leaking of Hillary Clintons emails illustrates this. Actors are looking to weaponize non-weapons to a degree that causes damage to their enemies, but while they themselves cannot be viewed as an aggressor in the conventional way.

The element of weaponizing anything that can be weaponized concerns also the security of the maritime industry. Hijackable ships could become a weapon in hybrid warfare. Regulations must be in place to secure maritime safety both for those at sea and for the rest of the world.

All scenarios above are theoretical. To this day no vessel has been digitally hijacked or manipulated to a degree where the shipowner had to pay a lot of money to avoid an environmental disaster. Nor has any ship been digitally hijacked by a government in the pursuit of getting the upper hand in the ongoing hybrid war. But then, if worst case scenario should pan out, even if there is a 0,0001% chance of exactly this scenario, is it worth the risk? Above several challenges to the introduction of the fully autonomous ship have been covered, and these illustrate that not one single solution will make the crew-less vessel possible on a global basis. Technology on its own will not be enough, nor will the introduction of necessary regulations in a single country or region. The complete implementation of autonomous sea going vessel requires a complete, global adaption to this situation to avoid any disasters, and to be sure that these vessels will make the world better and not compromise security and safety.