Fully Autonomous Cargo Ships Face Enormous Challenges

As with all “bleeding edge” technologies, the challenges are huge, and it’s possible they may end up derailing the whole venture.

By MC01, a frequent commenter on WOLF STREET:

In September 2017, a consortium of Norwegian firms, including Yara International, Vard, and Kongsberg, announced that they’d start testing the hull design of the YARA Birkenland at a test-tank facility in Trondheim. When finished, it will be a rather small ship – 3,200 Dead Weight Tonnage (DWT) – but it made big waves around the world: It will be the first fully autonomous ship to enter commercial service (image via Kongsberg Maritime).

Later this year, according to the timeline set by the developers and the regulators (led by the Norwegian Coastal Administration), the unfinished hull of the YARA Birkenland will be towed from the Vard shipyard in Braila (Romania) to the Vard facilities in Brevik (Norway) to be completed. The shakedown cruise and first testing of autonomous capabilities are scheduled to start in 2019. In 2020, the ship will be delivered as a manned cargo, with more capabilities being gradually transferred to the shipboard systems until fully autonomous capacity is reached in 2022.

The ship will feature a superstructure to house the crew which will be removed when regulators clear it for completely autonomous operations, further adding to “the future as seen from the past” look of the ship.

That’s the plan, and it’s a bold one.

The ship will be fully electric and will feature many advanced solutions, such as fully automated loading and unloading equipment, a permanent ballast (consisting of the batteries), and automated mooring systems. In short, if everything goes according to plan, this ship will be able to carry cargo between three Yara facilities along Southern Norway without human intervention.

While most of these features have been trialed before, this is the first time they are not just integrated in a single ship but put into revenue-generating service and hence expected to perform reliably in a real-world environment.

The all-electric propulsion system was chosen following negotiations with the Norwegian government. The first design choice was to use liquefied natural gas (LNG) as fuel (either in a gas turbine or two-stroke diesel engine), but regulators thought this was going too far. More importantly, the use of an electric propulsion system allowed Yara to access funding from the Norwegian government which will go towards covering part of the development costs.

So far, the project has been subsidized to the tune of NOK133 million ($15.7 million) but more funding will be made available in the future.

As is the case with all projects filled with “bleeding edge” technology, there are bound to be problems, and it’s even possible these problems may end up derailing the whole venture. If this happens, remember an old saying: You can recognize the pioneers by the arrows stuck in their backs.

But even before the YARA Birkenland scale model had been shipped to Trondheim to start testing, the whole world was whipped into what I call “autonomous ship frenzy,” with symposia being organized and widely advertised, chiefly to pitch the technology to potential investors and analysts, and stories full of hyperboles were predicting whole fleets of autonomous ships sailing the oceans in a few years.

Less advertised has been the formation of a Japanese consortium sponsored by the Japanese government and led by Mitsui O.S.K. Lines (MOL), which includes fellow Mitsui companies such as Akishima Laboratories and Mitsui Engineering & Shipbuilding, the Tokyo University of Marine Science, Nippon Kaiji Kyokai (ClassNK), and government agencies such as the National Institute of Port, Maritime and Aviation Technology, which will focus on a number of separate projects as part of the Ishin NEXT program, including:

  • A legislative framework for the operation of autonomous ships
  • Fully automated collision avoidance
  • Fully automated engine management
  • Coastal navigation assistance.

All these projects are to be integrated at a “later date” into the so-called “MOL Smartship,” a futuristic cargo ship which perhaps wisely MOL doesn’t advertise as being fully autonomous.

At the moment, this is by far the most ambitious program. Everybody agrees it’s bound to cost “a lot of money,” but nobody has any idea how much: As it often happens with government-sponsored programs in Japan, massive cost overruns are, if not exactly a given, at the very least expected.

However, several technologies stemming from autonomous ships will probably find their way into commercial operations far sooner.

Some of the technology Kongsberg is developing for the YARA Birkenland will be employed to assist, not replace, ferry crews on the Horten-Moss route starting in September 2018. Ferries are vital along Norway’s rugged coastline and innumerable isles, not merely to transport vehicles but also all kinds of goods, from foodstuff to construction equipment.

Kongsberg is purposely opaque on the exact nature of this technology, but it’s safe to assume this technology will make use of GPS/GLONASS data and LIDAR (LIght Detection And Ranging) to help mooring operations; and data from onboard sensors will contribute to autonomously regulate engines and improve fuel economy.

Robert Allan Ltd, a Canadian ship engineering firm specialized in harbor service vessels, is presently working on a “technology demonstrator” to test solutions which could be used on a future crewless fireboat.

Firefighting is dangerous business even with the best training and equipment. With LNG becoming more common aboard ships, both as a cargo and as a fuel, there will be a whole new set of challenges to the brave and hardy people manning fireboats and firefighting tugboats, or Fifi, as they are collectively and affectionately nicknamed in the business.

The Robert Allan concept does not call for a completely autonomous Fifi governed by computers. Though many systems will be self-governing, the fireboat will be remotely controlled by humans in a control center. The challenges are huge. Just think about providing the control center with effective situation awareness while an oil tanker is burning and half a dozen fireboats are dousing it with thousands of cubic meters of water. But the benefits could be equally huge, especially considering the boundaries between fireboats and harbor tugboats are very thin these days.

Robert Allan has also partnered with Rolls-Royce Marine to trial remote operations: A harbor tugboat owned by Svitzer (part of the A.P. Moller-Maersk group), which partly financed the test run, was modified to allow a single crew member to pilot the tugboat inside the Copenhagen harbor. In spite of the PR blitz, the test was very short and hasn’t been repeated yet: The tugboat undocked from the quay, turned 360°, passed in front of the Svitzer harbor offices for a photo session, and returned to the quay. Danish marine authorities allowed the test on conditions the tugboat carry a full crew and that autonomous mode could be manually deactivated at any moment. Again: no chances are taken here.

UK yacht manufacturer Hushcraft is developing the SEA-KIT technology demonstrator under a program financed by Royal Dutch Shell and by the Nippon Foundation (image via Hushcraft):

This vessel is designed as a mothership for an AUV (Autonomous Underwater Vehicle), a mini-submersible originally developed for military applications but which, given the tepid interest by navies worldwide, has been adapted for underwater mapping and exploration purposes. Royal Dutch Shell, a world leader in deep-water exploration, has a strong interest in deploying these vehicles to map seafloor features which may be used to narrow down the location of exploitable deep-water hydrocarbons. Search operations such as those undertaken to discover the wreckage of the ill-fated Malaysia Airlines flight 370 are also touted as another possible task.

These AUVs – or “robosubs” – go back to 1965, when six “fishes” as they were called, were ordered from Westinghouse by the US Navy for the first-generation nuclear submarine USS Halibut (SSGN-587), which was being covertly transformed into an intelligence gathering platform. Each of the hand-built fishes cost $5 million, an enormous sum in 1965, and had been covertly financed using “cost overruns” in other programs, most of them fictitious and created exclusively to leech funds for the Halibut conversion. Alas, as the Navy was understandably reluctant to lose any of them in action, they were tethered to the USS Halibut by a 7-mile-long steel cable.

A final word of caution. It will be many years before truly autonomous ships will take to the seas: Even if the YARA Birkenland project proceeds smoothly and according to schedule and regulators don’t get cold feet for one reason or another, the ship will be limited to a 12-nautical-mile coastal route while being remotely monitored by two or three control centers (the exact number is still to be decided). Any accident will negatively impact the whole venture; public opinion appears to be very sensitive about autonomous vehicles even as the hundreds of accidents involving ships with a human crew worldwide each year seem to fly below the radar.

Specialized semi-autonomous vessels, such as remotely controlled fireboats, may enter limited service in a relatively shorter time, but one always needs to remember what ultimately drives automation: money. As long as humans can do a given job for less money than a machine, automation remains in the realm of science fiction. By MC01, a frequent commenter on WOLF STREET.

In the iron-ore trade, everything revolves around ships and China. Read…  The Largest Ships in the Huge Iron-Ore Trade

Enjoy reading WOLF STREET and want to support it? You can donate. I appreciate it immensely. Click on the beer and iced-tea mug to find out how:

Would you like to be notified via email when WOLF STREET publishes a new article? Sign up here.

  82 comments for “Fully Autonomous Cargo Ships Face Enormous Challenges

  1. endeavor says:

    Who will be there to repel the pirates in mid ocean?

    • Ethan in NoVA says:

      The pirates can use GPS spoofing hardware and huge electromagnet (to simulate north) and get the autonomous ship to follow them home.

      • elysianfield says:

        ” huge electromagnet (to simulate north) and get the autonomous ship to follow them home.”

        Yup, And they can order one today from the “Acme Electromagnet Company”….

    • Nicko says:

      The unfolding and burgeoning autonomous revolution will define the 21st century.

      • Kenny Logoffs says:

        As a huge failure.

        Sandbox efforts look so promising.

        Show me a truly competent AI or autonomous car.
        Ie, can even work in such harsh conditions as light rain.
        Or leafy damp road with puddles!

        The only revolution will be in jobs where humans are already wasting their lives doing them, and are being phased out by automation generally already and into the future.

        None of these truly autonomous tools will work better than humans generally until AI is better than humans generally.
        When that happens everything will become irrelevant, including humans.

        • Wolf Richter says:

          Humans are terrible driers. Human drivers are responsible for 36,000 deaths on US roads and millions of seriously injured. It doesn’t take a whole lot to be better than these human drivers.

    • Rates says:

      Amazon Mechanical Turks.

      Who else?

      • Rates says:

        Or TaskRabbits or Fiverrs. Doordashers though will weep because business in the open sea will go from zero to zero.

    • Finster says:

      Autonomous ships will be dealing with autonomous pirates.

    • MC01 says:

      Piracy these days is almost exclusively a problem in coastal waters.
      While the Indian Ocean gets all the glory (and the military action), it has never been as problematic as the Straits of Malacca and the Gulf of Guinea. I’ve read the waters off Venezuela are becoming pretty dangerous as well lately.

      Modern day pirates, differently from their predecessors, are rarely interested in the cargo. They chiefly aim to rob the crew (most ships carry a relatively large quantity of cash, at least by Third World standards) and steal any supplies they can sell easily. That’s why fishing boats are targeted so often: their equipment is easy to resell. Kidnapping the crew, to demand a ransom from the ship owners, while well documented and well publicized, is ironically enough more of a problem in the Gulf of Guinea than off the Horn of Africa.

      We are still likely decades away from truly autonomous ships, as I said, so that’s a problem for future regulators to tackle.
      As a Chinese academic said: “We have to leave something for future generations to do”. ;-)

    • Patti Seibold says:

      That’s exactly what I was thinking while reading this. More Pirates on the Seas

  2. unit472 says:

    Interesting but what about high tech pirates or just the old fashioned ones stealing an unmanned ship?

    • Wolf Richter says:

      Like stealing a manned ship? Happens all the time. They take the crew hostage for ransom. With an unmanned ship, there’s no crew, and no hostages, and no ransom. So the incentives go away.

      • gary says:

        As much as I am skeptical of autonomous driving, it could actually work well with ships.

        The reasoning is obvious: much, much less interaction with others in the middle of an ocean (and the few interactions there may be can be handled with radar, algorithms, etc.)

        Furthermore, once closer to shore, the ship can be taken over by humans in some manner (remote control, etc.)

      • Ensign Nemo says:

        What about terrorism?

        Iran claimed that they ‘hacked’ a US drone and made it land in Iran. This claim is dubious, but it shows at least that the bad guys are thinking about such things.

        What if one of these things was hacked and then deliberately scuttled in a narrow strait, such as Hormuz or the Bosporus?

        How much damage would a supertanker full of oil slamming into a city’s waterfront do? How much money would be spent to clean up the mess?

        I can’t imagine that they would ever let a LNG vessel become unmanned, those could be converted into floating bombs. There is one near Boston Harbor:

        “While scientists say an LNG tanker is highly unlikely to explode in a fireball if hit by a bomb or missle, much is unknown about how a large cloud of gas fed by slowly vaporizing LNG might ignite and what that could do to the densely packed waterfront neighborhoods of downtown Boston, Charlestown, East Boston, and Chelsea. ”


        • Wolf Richter says:

          Hacking is a pandemic and is hitting every device, and every device maker is fighting back. Cat and mouse. Some folks are trying to hack into nuclear power plants. The effects would be worse than someone hacking into a tanker.

        • Joan of Arc says:

          How do they recharge the batteries? Are there solar panels on the ship?

      • Franco says:

        There is a whole automous ship and its cargo to hold for ransom.

        • Wolf Richter says:

          Yeah, try to sell the cargo of a ship on the black market…

          And how are the pirates even going to control an autonomous ship?

          And if the only people on that ship are pirates, and no one else, one of the interested military forces will send a missile into that ship and videotape what happens to the pirates on that ship. The video will go viral, and that’ll be the end of trying to take a crew-less ship hostage.

    • Prairies says:

      If I were tech savy enough to be a modern day pirate I would not steal the whole ship. There is no crew, and security is on land waiting for alarms. So I would steal expensive tech or even disable the GPS so it is left adrift at sea, be a real James Bond villain and use a submarine to get away lol.

  3. nick kelly says:

    As long as the bleeding at the ‘bleeding edge’ is not literally blood.

    • 91B20 1stCav says:

      Silly holiday-weekend idiomatic question here: when did ‘leading edge’ become ‘bleeding edge’? (My guess is about the same time ‘home in’ became ‘hone in’. Perhaps the leading edge was honed to a bleeding one?). A happy Labor Day to all.

      • MC01 says:

        It’s a term I first heard from an old systems designer who had worked at the Autonetics division of the North American Aircraft Corporation (now part of Boeing).
        Apparently somebody made a joke about the fact that working on ultra-advanced programs (solid state data analyzers: these were the late 50’s/early 60’s) with very aggressive specifications and very demanding deadlines was akin to performing a balancing act on a razor edge and since those programs turned out to be a smashing success, the joke became that whereas companies such as Lockheed and Westinghouse worked on “leading edge” technology, Autonetics worked on “bleeding edge” technology, as standing on that razor edge requires cutting oneself.

        Kinda like the origins of the name of the H2S radar, it’s an obscure joke, but I really like the sound of it, as it gives a good idea of the sweat, tears and blood that are expended on ultra-advanced programs.

        • 91B20 1stCav (AUS) says:

          MC01-a pleasure-to-read article (as always), AND a delightful explanation for an idiom that I believed to be the child of inattentive hearing and our general fascination with that which usually appeared on a newspaper’s front page. Now, if I could only figure out how we went from honing something OUT, or TO an edge, my weekend would be complete! (sorry for the caps, Wolf). Again, a happy holiday to all.

        • MC01 says:

          And thanks to you as well.

      • nick kelly says:

        In general the fact that truly novel systems can have all kinds of unforseens, huge costs, etc. etc.

        It happens a lot that the person or co breaking the trail is NOT the one who turns it into a commercial success.

        E.g. the worlds first jet airliner was the UK’s DeHavilland Comet but it was also the first one to ‘discover’ metal fatigue and after two were lost the program lost its lead, as did the UK in jets.

  4. paul maier says:

    once the hackers get involved bringing things to a
    standstill, this will not be as “glorious” an idea.

  5. Earl Smith says:

    This is really not so new. I recall back in 1966 when my submarine (on the surface) had to dodge a freighter that was sailing in the Florida Strait south of Key West. Our problem, there was NO ONE ON THE BRIDGE of the merchantman. Even while on the surface the submarine has the visibility of a small car and would be ignored by any crew on the ship, if any were actually there.

    Just one more hazard for naval officers to be aware of.

    • MC01 says:

      Collision avoidance technology has advanced by leaps and bounds since the early 70’s. However this doesn’t mean the best systems and protocols are universally adopted: there are still far too many ships sailing with only the most primitive collision avoidance systems and with crews trained only with the most basic protocols.

      To give an example most offshore platforms around the world right now have not just collision avoidance systems “wired” to automatically alert approaching vessels through both radio messages and visual/acoustic signals, but those same systems can also warn the platform crew of the presence of unauthorized small crafts near the platform.
      The exploitation of the Hibernia oil field (roughly 200 miles off St. John’s) even required the creation of an early-warning system to spot icebergs from at least 20 miles away and plot their course so that platforms can be evacuated in case of a possible collision: the fog banks hanging over the area most of the year make visual tracking impossible. It’s a very slick and highly impressive system, even for somebody with limited IT knowledge as me.

      As it often happens the technology is there: it’s a matter of making it useful.

      • fajensen says:

        Heh! When I worked with RADAR design, the joke was that the only enforced requirement for a marine RADAR was that the antenna would rotate when the ship was docked!

        We proposed a phased-array type RADAR with frequency diversity (using and compensating for doppler-effects) for the newly introduced fast ferries. A RADAR quite similar to the kind that equally fast-moving battle-cruisers use, but, COTS and much, much cheaper. Nah. No Good!! Went with the El-Chepo magnetron type.

        Remember that, when sailing into the path of one of these things: They can’t see much even on a good day and the entire image is skewed because of processing delays that are irrelevant at normal ship speeds.

        Nowadays the solid-state RADARS with advanced signal processing is preferred (and cheapest too because they are 99% the same as what goes into a mobile telephony antenna system) so we were indeed right, we were just a bit too early in relation to the decision makers retirement schedules.


  6. scotts71 says:

    Great article, again. However, I cant think how these, marginal, staff cuts, as a percentage of opp cost can drive exciting results. From what I have read by you, its more of an issue of over. Does the ROI work?


    • MC01 says:

      The Ishin NEXT program I wrote about in the article will give us an idea of what return on investment (ROI) a large shipping company with worldwide operations such as Mitsui O.S.K. Line can expect from automation: the YARA Birkenland is more a small scale project dictated by peculiar infrastuctural and labor issues in Norway and, albeit very useful for development, is a poor meter for worldwide, deep water operations.

      Bear in mind that in just five years automating many ship systems, working seriously towards improving on-board systems reliability and implementing new corrosion inhibiting technologies (thus reducing on-board maintenance) have allowed to dramatically cut crew sizes: the CMA CGM Explorer Class (first delivery in 2009, 16,000TEU) have a crew of 25 while the Maersk Triple-E Class (first delivery in 2013, 18,000TEU) have a crew of just 13. Both classes also have accomodation for two cadets, albeit these are not needed to operate the ship at full efficiency.
      To give an example the four even larger (over 19,000TEU) China Shipping Line (CSCL) container carriers that have entered service since 2014 have a far larger crew of 31.

      Intriguingly enough the Triple-E with their skeleton crew have a better safety record than the CSCL behemoths: the CSCL Indian Ocean was grounded in the Elbe River in 2016. According to CSCL this was due to a “rudder control malfunction” but German authorities have since put the blame on old fashioned human error. As can be imagined freeing such a ship, which had grounded herself at high tide, was no mean feat and required a small armada of tugboats and two dredgers.

      • scotts71 says:

        MC01, thanks for responding, appreciated. I did some digging around and was able to find that respondents to an operating survey in 2015 (https://maritime-executive.com/article/ship-operating-costs-set-to-rise#gs.LDO8UQ ) crew staff cost as percent of operations was 20-60%
        (which highlights your comment). However, what is interesting is, that crew staff is from ”low cost wage areas”, ie chinese crew sailing to the US. So %s are %s, however, what is it in terms $ per voyage? divided by container or CBM of goods? Must be small, right.

        Interesting about a ships running aground, as I would have thought GPS and advanced systems would prevent this, but maybe they dont have them currently?

        Anyhow, it seems oversupply is the driving force for this commodity business collapse. But at least they secured financing to take on more debt to buy new boats, expand dry docks, buy other shipping lines. If you borrow money at, 1%, surely, you can find away to make it work. Like you said, you still need buyers for your cabbage. If everyone has autonomous, cabbage pickers, you, and they, can drop your price and you sell x more cabbages, untill, you have reached max cabbage output/consumption at the lowest cost basi. Then where do you go? As autonomous machines dont eat cabbage!, people do.

        • MC01 says:

          Crew is about 40% of average cost of operations nowadays. It’s not such a big concern right now as bunker fuel, which has been suffering from serious fluctuations in prime markets such as ARA this year, but it still something you need to keep an eye on.

          AP Moller-Maersk has been opting for smaller and better trained crews (those systems may be automated, but they still need somebody to keep an eye on them) backed by strong investments to reduce workload and on-board maintenance, and that’s probably the way forward right now.
          Chinese shipping companies such as COSCO and CSCL seem to be still opting for larger crews and lower investments, but I suspect they’ll be forced to embrace automation as well, for the simple reason you cannot sail or moor such enormous ships without a lot of help from computers and software. A few more accidents such as that of the CSCL Indian Ocean and they’ll be blacklisted from ports, like it happened with the STX-built Valemax ore carriers.

          What makes Chinese shipping companies such formidable competitors however, are not labor costs but their ability to access a capital market that for all purposes is a complete mystery to outsiders.
          How much does COSCO spend every quarter to service their obligations? How much of China Shipping Lines’ cashflow needs to be diverted to pay out bondholders every FY? The answer is a resounding “we have no clue”: the numbers these companies deign to divulge fall somwhere between “useless” and “complete fabrications”.
          This doesn’t mean these Chinese companies have access to unlimited funds whenever and whatever they want: it only means the Chinese capital market is so jumbled, so contorted, so willingly opaque and convoluted we literally have no idea how feasible these shipping companies are.

      • fajensen says:

        Intriguingly enough the Triple-E with their skeleton crew have a better safety record than the CSCL behemoth

        Maersk has possibly the best trained crews on the planet and they have full authority while on duty; There is an “Aura of Command & Competence” about them all, its quite something to see.

        Maersk leadership has come up “the hard way” through delivering results within the Maersk organisation, the culture is that they do not hire “outsiders” directly into positions of significant authority.

        If one is very talented and like a real challenge, one can perhaps get into a somewhat higher starting position by going through “Maersk University”, which is the colloquial name for the various candidate and intern programmes offered to “newcomers”.


  7. scotts71 says:

    Sorry, …..”over supply”

  8. Iindie says:

    At last ! Autonomous tugboats and barges will allow for much faster logistics to island nations as well as archipelagos especially if the power to the engine is green ! Automation will surely help since number of capable cargo sailors has been crashing to multiple decades low , getting crews for long runs is becoming a headache and it will not get better in the long run . Crews for cargo is the main driver for increase of tonnage per boat we are still seeing these days , automation will put back crews where it matters! At this point no one wants the job , the technicals and training versus pay does not add up , so automation is the only way the industry can rebalance . So the only way forward is to automate and go green since shipping companies are not making any effort to make the career attractive or liveable 35k USD / year !

    • MC01 says:

      You are right: getting certified crews these days is becoming a headache, for no other reason wages aren’t exactly attractive, especially on bulk carriers.
      That’s the reason so many shipowners are ready to close both eyes and turn the other way when hiring people.

    • Olivier says:

      See? This is the irrationality I was referring to earlier (but below). This is a simple problem that could be solved very easily by offering better pay but the class warriors in the executive room would never consider it: for one thing it would ruin their class warrior creds with their peers. Can’t have that. Thus instead come the robots, even if their development will cost billions.

  9. Paulo says:

    Autonomous oil tankers? I don’t think so.

    Just how thin are the profit margins, anyway? When I occasionally dropped off ships pilots and customs agents on freighters (BC Coast) I used to snoop around. Officers were always British or Japanese and the crew were Phillipino. Are you telling readers companies plan to not even run a skeleton crew for emergencies? There are hardly any crew on these ships, anyway. They paid the Philipinos sweet f all. Sometimes, crew don’t even get paid and the ships gets anchored up and idled.

    Electric? Powered by Batts? Range? Come on, they can’t even make an EV with a 300 mile return range let alone a globe trotting freighter. Windmills and solar to recharge? Not likely. Electric assist sails to augment conventional powered ships I get.

    Self loading? Not in Canadian ports. The International Longshoremans Union might have something to say about that. You think an insurance company will insure cargo and liability without crew standing by? I don’t. I can’t see any large city allowing autonomous ships in their harbours. Hell, The US navy just had two major accidents and that’s with all the bells and whistles and command structure imaginable. They even had scapegoats!

    This is taking automation too far. People need jobs and careers or society will tear itself to pieces.

    • Iindie says:

      Freighters and tugs are where the main interest would be I think . Tankers are much too risk prone ! And solar does make a LOT of sense for coastal workloads . In calm waters where sunshine is no issue , not British Columbia of course where latitude and weather conditions renders the use of such technologies impossible or inefficient at best !

    • Rates says:

      Society will be reformatted. Once robots are fully functional, the owners will use them to wipe out the serfs, leaving them the masters of the earth.

      “People need jobs”. Wipe out the people. Society can’t tear itself apart by definition.

      Huge pension liabilities? Wipe out the pensioners.

      See the similarities? There’s always a solution to every problem.

      Final lesson: Be the 1%. Better still be the 0.1%. Can’t beat them? Join them

      I am not advocating the above, but you can just sense the thought patterns of our elites.

      • briny says:

        I’ve noticed the same trend line. Vaguely related, but ever notice the number of prescription drugs, derived from transplant-rejection prevention medications, that are entering the mainstream as treatments today? One good flu pandemic is going to star in a future production on that.

    • Joan of Arc says:


      That is what I am thinking. What is the process for charging immense banks of batteries? Will solar panels aboard ship help? What is the range of the ship per full charge? The batteries could provide power by night and if they line the upper deck with solar panels they could help recharge the batteries on sunny days. At port will they use 660 volt superchargers? Will they ask Tesla to help so that Elon doesn’t tweet and call them a bunch of “p_dophiles”? LOL

  10. Pat McKim says:

    This is a silly concept. And no it’s not just about costs. It’s about safety. If a ship hits another ship and many are killed and property damaged, its a huge cost.

    For anyone who has stood a bridge watch in a rough sea at night, while transiting through a choke point (straights) with casualties in some systems know it’s obvious why this doesn’t work.

    There are too many unknowns, unplanned unknowns and worse reverse feedback…. Things not charted under the water like shifting sand bars, ships changing course unexpectedly, sea states changing, wind changing, tides changing, and most importantly casualties on your own ship. I could go on. The sea is a most hostile environment. Murphy’s Law was born on the sea. All kinds of things go wrong, very wrong: casualties to equipment, electronics, floodings, fires to name a few. Salt air screws up electronics, salt water also screws up machinery. Sensors goes dead or worse give opposite feedback–it happens, just as it does in the air and caused the Air France crash from Brazil to France. Machines and electronics don’t have common sense; they don’t a seaman’s eye that takes years of watch’s at sea and years of study to gain. Sitting on a rail in an around the bouys race ain’t going to sea. Minimizing crews makes sense, but eliminating them? The Navy has just had problems with collisions in choke points because they weren’t trained and focused enough. I doubt that having some system with limited experience with software coders that can’t anticipate everything would do better and most likely worse. I was in tech, internet software and defense technology, as well at years at sea on warships and big boats. There are places where it makes sense and places where it doesn’t. As usual tech idiots want to take it too far.

    • Wolf Richter says:

      Pat McKim,

      I just love this – to pan the technology of autonomous ships because of this, as you said: “If a ship hits another ship and many are killed and property damaged, its a huge cost.” You’re clearly clueless. Ships hit ships all the time, including two recent collisions between cargo ships and US Navy ships — all of them manned by humans, including naval officers.

      “As usual tech idiots want to take it too far,” you said. This is funny, given your comment. I think you’re confused as to who the “tech idiot” is.

  11. 4corners says:

    I understand the desire to innovate in propulsion and navigation, but up to the last paragraph of the article I was wondering what the impetus was to go crewless. It would seem the cost of the crew is negligible relative to the capital and operating costs of a freight vessel. Is that not the case? Are the margins so thin to justify going bleeding edge? There must be a reason these companies want to be early adopters of autonomous technologies rather than simply incorporating them once bugs are worked out.

    • MC01 says:

      In a country such as Norway autonomous ships make a lot of sense as not only labor is generally very expensive, but specialized crew roles (helmsman, navigator etc) are even more so as the supply is severely restricted, a rare case of unions doing their job when it comes to immigration policies. Multiply that for the hundreds of ferries and coastal cargoes in operation at any moment.
      That’s why the push towards automation is so strong there.

      Generally speaking, and this applies to all industries, automation is highly palatable because it allows not just to do away with human error (think about what Wolf always says “humans are terrible drivers”: that’s true, no matter how much that can bruise a lot of egos) but have consistent performances.
      In the maritime trade consistent performances chiefly means one thing: reducing fuel consumption and engine wear. That alone makes the idea of autonomous ships palatable. And if they prove even marginally safer than manned ones, insurance costs are bound to go down as well.
      Do I need to say more?

      • 4corners says:

        Thanks for that explanation. On the second point, it will be interesting to keep score on human error versus machine error. From the little I know about autonomous driving, the conditions of the road are critical. I can only imagine the unexpected situations that arise at sea.

  12. nick kelly says:

    A solution in search of a problem.
    And very likely to create more problems than it solves.

  13. EcuadorExpat says:

    Jets on navy carriers have the ability to land without pilot intervention even in the roughest seas. Compared to this automating a ship is child’s play.

    My guess is the developers will spend more time fighting the old guard than they will developing the systems. From a systems developer point of view, this is control technology that would easily scale up, so it can be developed cheaply on a small scale.

    Like eliminating the pilots from fighter aircraft, or much simpler, having weapons that fire to the rear of such an aircraft. “He’s on your Tail!” “Great!” Pushes fire button.

    Many small towns in the USA died because the railroads never developed a way to separate/add a car (or load) to a moving train. Every small town required stopping, breaking the string and moving cars to a siding, retrieving empty cars and reconnecting the string. This often consumed more fuel than actually moving the loads between two destinations. Today a smart robotics high school student could figure out how to do this.

    • nick kelly says:

      One difference: the carrier has beacons in communication with the aircraft.
      Objects at sea, freighters etc. won’t.

      Second difference: money for the plane/ carrier interface is no object.
      For the commercial ship, money is THE object.

      Re: eliminating pilots from fighters: unless the plane is an automaton and not a robot, it would easily have its coms jammed.

      Re: jamming. The US Army has publicly released some conclusions from the war in Ukraine. The Russians have deployed a cheap, portable, low- power transmitter that jams GPS for about twenty miles.

      One Army response: it has started some special forces training in the use of star charts for night time navigation.

    • Frederick says:

      They seem to have perfected that auto/train connection pretty well in Switzerland with the train tunnels under the Alps But that IS Switzerland

      • MC01 says:

        Have you traveled in Switzerland in the last few years?
        The traffic is getting simply ridiculous pretty much everywhere.

        This is not the place for a full discussion, but suffice to say Switzerland is a whole lot like Japan: infrastructures are basically make-work projects for local construction groups and GDP boosters, only they run much, much longer.
        The roadworks on the 19 and 27 Cantonal roads in Graubunden have become the stuff of legend, but I can announce that after eight years of mostly doing nothing the longest-running one (opened in 2010!) seems close to completion. Don’t get your hopes too high however as the Canton announced two new roadworks on the same roads that will run at least to 2023!

    • MC01 says:

      They are two completely different problems: you cannot really compare the thousands of parameters the data analyzers running even an autonomous dinghy would need to elaborate with the two way communication between aircraft and land that takes place through microwave transmitters and receivers.

      I’ll also be sure to tell my acquaintances working in logistics the high school students in your area know how to separate cars from a moving train without causing complete carnage. It seems for all our technological prowess that kinetics still apply, but I haven’t studied physics since the Uni so I may be out of the loop.

  14. Mike G says:

    One of the jobs of the small crews on container ships is grinding off rust and repainting. Unless autonomous ships can be made rust-proof they’re still going to need crews. Likewise for maintenance requirements for the ship, engines, container reefer units and the like.
    Crews are a small component of the cost of operating a container ship at this point, so I can’t see this being economical.

    • briny says:

      Ex-career US Navy here and that’s at least half the reason to have a crew, dealing with all forms of corrosion at sea. You chip and paint from stem to stern and once you are done, you start at the stem again. The other part is having people on hand for when things go wrong. My service has gone to automating as much as possible and it becomes a serious problem without a resolution yet in the event of collision (USS Lake Champlain, USS John D. McCain, USS Fitzgerald for most recent).

      You really need that crew available for damage control. What’s been happening on the LCS and DDG-1000 don’t exactly inspire my confidence for the US Navy’s future. Navigation and damage control featured prominently outside my engineering role, so I’ve some background to give bones to my worries.

      For the autonomous aspect (AI/ML are areas of interest), getting rid of the people who do the damndest things could be a plus, when things are going right. Remember, there are at least two actors in any collision at sea. So, half of the problem perhaps solved, perhaps not. And I’ve no idea on how Lloyds is going to go about insurance for these beasts. Really, it’s going to come down to monitoring risks to determine whether such a market segment is worthwhile too.

    • MC01 says:

      The Maersk Triple-E container carriers have a crew of 13, or about half the complement of the previous record holders, the CMA-CGM Explorer Class (25 crew). Besides there’s always room for two cadets on board, albeit these days the positions are rarely filled.
      There isn’t really a whole lot of maintenance that can be done with such small crews, especially given the need to monitor ship systems dramatically curtails the number of people you can send around the ship to do maintenance.

      Older ships, however, are another matter completely: there are plenty, and I am not using this word lightly, of bulk carriers and container ships built in Japan in the early 80’s still in service, especially on routes where MARPOL protocols do not apply or are not enforced.
      Life on these ships is basically maintenance and repairs first, and sailing second.
      Kinda like when I worked on my grandfather’s farm: you spent most of your day fixing crumbling equipment and only a minority of it working in the fields. :-)

  15. Olivier says:

    Thanks for answering my wish. Re. incentives, our entire modern civilization is based on the worship of technology, with “solutions” always presumed to be better the more technology they employ. Thus for all the talk of companies being profit-driven you should never discount the element of irrationality whenever “advanced” technology is involved.

  16. Cashboy says:

    As a simple accountant, I would say that the cost of a crew on a large cargo ship is a very small cost in relation to the running cost (fuel and on going maintenance) of a ship. Most of the maintenance of a ship is generally under taken when a boat is in service by crew members so can’t see the accomodatio part ever being removed.
    Regarding pirates, the pirates just need to become techy (like modern day bank robbers) infiltrate the computer servers and take control of the boat and take it to wherever they want or threaten to ram it into rocks or switch of sump pumps and flood the batteries making the ship redundant.

  17. nick kelly says:

    My dad ran a commercial fuel barge for years and I’ve sat around with tug boat guys a lot.
    To me tugs seems like the LAST candidate for going unmanned.

    The majority of tug work is short haul. This is true even where there is load (barge) but even more so when pushing a large ship out of dock.

    The guys around here were often towing barges. Half the job is attaching the cable, unhooking it and watching it and the barge in between those two. Steering with a load in a wind is completely different.

    Most tug work is done in congested waters, not out in the deep lonely blue where the course is the same for days.

  18. wkevinw says:

    Last mile is the “problem”. For the long haul, open ocean (and, e.g. open road), robotic pilot seems feasible. Or at least you can “train” ships or trucks with one pilot for multiple vehicles or boats; cutting costs to employ people.

    The complexities of congested areas and small objects are difficult to automate.

  19. John Taylor says:

    “As long as humans can do a given job for less money than a machine, automation remains in the realm of science fiction.“

    There was a period in the Roman Empire when slave labor started to become more scarce, and you started seeing an explosion in labor-saving technologies such as water wheels and windmills.

    Similarly there was a period when Switzerland was known for their mechanical devices (like the famous Swiss watches) as they reacted to their relatively expensive labor with more creative mechanical labor-saving and production-enhancing devices in industry.

    Labor, after taxes and regulations, is still expensive enough to encourage the research of labor-saving technologies – but the main interest now is in adjusting to the variability of demand.

    There is a boom in freight demand which the industry is incentivized to meet, but they don’t expect it to last. It’s easier to turn off machines as things slow down than downsize the labor force, but more importantly it is a lot easier to bring machines back on when needed than to re-build a labor force which was gutted in the last slowdown.

  20. Bookdoc says:

    my question is about insurance. I wonder who will be the first to offer liability on an autonomous ocean going ship. I am already hearing about insurance issues with the tesla. The body shop at my old dealership said there’s no way they would even touch a tesla in the body shop. That’s without fully autonomous driving. Think about what the first accident one of these ships gets into will do. I do believe, knowing a little about the nature of the ocean, that autonomous ships will not be able to deal with unusual weather conditions just like autonomous cars haven’t been shown yet that they can deal with a typical northeastern winter.

    • MC01 says:

      The issue with Tesla is the lack of spare parts and technical literature. Basically the same issues as with any luxury car not made by a serious automaker such as BMW or Toyota. That’s why bodyshops are so reluctant to work on these cars, whatever the brand.

      Mitsui O.S.K. Line has very close ties to the Mitsui-Sumitomo Insurance Group (the product of a 2001 merger between the insurance arms of the Mitsui and Sumitomo keiretsu), and I am sure at some point in the future they will be brought in.
      As I wrote above, however, ships with a high degree of system automation such as the Maersk Triple-E class have shown themselves to be as safe or even safer than ships of the same size with less automation. As the data keep on pouring into the Lloyd’s database I am sure we are bound to see some pretty nice surprises. ;-)

      • John Taylor says:

        The problem with insuring autonomous vehicles and ships isn’t about safety records, it’s about assigning blame.
        If one incident happens with an autonomous ship or vehicle, it’s certainly plausible that all financial blame would hit the manufacturer and potentially bankrupt them. You can imagine all the costs from property damage, environmental damage, cargo damage, etc. that they could pin on one company.
        Back in the ‘90’s, any boat with a radar system hitting a boat without one was automatically at fault by law, so my friends father said he’d never get one.
        That being said, most of the cargo ships don’t flag in the US and spend the bulk of their time in international waters, so they would probably flag somewhere with more favorable litigation systems anyway. Autonomous cars and trucks will have a more difficult time with that.

  21. Wisil says:

    Ranking autonomous vehicles on degree of complexity of required control systems:


    Trains are clearly the easiest, but labor pushback has been strong.

    I think ships are a logical choice, but remote control via a satellite uplink will come first. Given enough sensors, such as LIDAR, infrared and CCTV, you could have one captain controlling several ships via a remote location. Piracy would not be a problem, since the helm can be locked, and the boat steered to the nearest navy vessel. GPS spoofing would not work, since there are systems that immune to jamming or spoofing. Autonomous will happen, but there will always be a remote control backup if needed.

    • nick kelly says:

      There are no systems relying on EM coms like GPS etc. completely immune to jamming.

      Jamming goes back at least to WWII and apart from pure military jamming there has been an ongoing jamming war between political enemies. Voice of America was always being jammed during the Cold War.
      The adversary has to detect the frequency you are using (scanning) and then transmit on that frequency.

      Once you take to the ‘airwaves’ i.e., radio (EM waves) they are public. You can’t stop the guy listening although you can try to encipher your message. Of course if he is trying to find out WHAT your message is, the last thing he will do is jam your signal.

      One result of Russian jamming of GPS in the Ukraine conflict is switching
      back to star charts for night time navs.

      The only solution invented in WWII is to constantly change frequencies, but this is not practical for real- time uninterrupted control.

      • nick kelly says:

        PS: One might expect that undersea cables would have been abandoned by now, but they their relative immunity to being snooped compared to EM waves of any frequency keeps them in ‘business’, just not commercial business.
        Their immunity is relative. During the Cold War the US tapped the Soviet undersea cable in the Ivy Bells operation.

      • Wisil says:


        Inertial navigation with GPS. Redundancy, jam-proof, and problem solved. The minute GPS jamming starts, the inertial system takes over and turns the ship to the closest available navy ship.

        There are even more advanced systems that don’t rely on RF signals or inertial navigation, but I would have to kill you.

        • nick kelly says:

          The Navy has better things to do than run around chasing Nav failures on commercial shipping, from whatever cause.

          ANYTHING is possible. Much less is practical.

    • MC01 says:

      Rio Tinto has already started to use a fully autonomous train on revenue-generating routes in July.
      Unions are mumbling, but they can do little about it, as many systems have already been fully automated on all Pilbara railway networks: even the most alert and experienced drivers need all the help they can get with trains that size.

  22. Mean Chicken says:

    Fascinating article, I wasn’t aware battery powered ships are under development.

    • MC01 says:

      The technology itself is nothing fancy and the concept is actually pretty old: electric boats were actually all the rage before WWI.
      However back then electric boats were almost exclusively used on passenger services: think launches shuttling passengers from and to cruise ships or small ferries. Submarines came later, when J.P. Holland had the idea of applying electric propulsion to them.

      The most innovative concept is the idea of using the battery as a permanent ballast on such a large ship but again the idea is nothing particularly new and has been trialed on progressively larger designs, mostly small ferries. That’s what fascinates me the most about the idea.

    • Wolf Richter says:

      Also don’t forget submarines. All except nuclear submarines switch to electric propulsion and battery power when they dive. This technology goes back over 100 years.

      • nick kelly says:

        The lead acid battery is one the oldest techs still doing the job. Getting close to 140 years old I believe.

  23. GOF says:

    Is there nothing sacred anymore? How can that future artistic wonder, writer of magnificent tales of adventure, generate any inspiration when they can’t run away to sail the seven sea’s. Where will the compasioate sea cook be, the brutal boatswain or the indifferent officers? No doubt stowing away will be dead simple with a crewless ship but how would one live without water and working in the galley? We are taking all the opportunities for adventure away from future generations.
    On a serious note, one of the unwritten commandments of maritime tradition is the obligation to go to the aid of another vessel in distress when life is in danger. This requires manpower in the final stages of rescue. How does an autonomous ship go to the rescue of a cruise ship or a capsized sailboat?
    One last point, and I admit to scim reading most of the replies, regardless of the writers slant on marine things. The examples cited by Pat McKim are very valid – other than the cheap shot in the final sentence and to which I take exception as one of the techies involved with automomous unmanned aircraft. Until a computer can assimilate and evaluate a myriad of variables in the context of human experience with the speed of the human brain, formulate a response and then execute it, it will remain what it is; a limited function machine devoid of compassion.

    • Mean Chicken says:

      “Where will the compasioate sea cook be?”

      She goes by the name of Rosie! :)

  24. Roland says:

    What’s this about “automated mooring systems” ? Hand-waving. Robots throwing heaving lines, threading mooring lines thru the hawser ports, fitting rat guards, etc.? Can a robot handle a boathook?

    Big ships need synthetic super-ropes. Very expensive. This justifies having a winch for each rope, which has reduced the need for people. But I don’t see a next step in this area.

Comments are closed.