Samsung floats autonomous ships as ready to sail in 2022 Samsung has floated the prospect of soon selling autonomous ships, if it can sail one successfully in August 2021. The Chaebol says it’s teamed with Korea’s Mokpo National Maritime University to fit the Samsung Autonomous Ship system to a 133m-long, 9,200 tonne training vessel and sail it from the port of Mokpo to the island …
How about something along the lines of...
https://en.wikipedia.org/wiki/Phalanx_CIWS
Let's hope there are no biases on the AI training material. Not using footage from The Pirates of the Caribbean would be a start
You need two ships, then if one of them is taken over by terrorists you can use the other ship to colonise the islands, gather resources and upgrade your ship, and hopefully take out the other ship.
Er..
It's much easier to handle pirates on an autonomous vessel. Firstly it could be made very difficult to gain entry, with all external hatches welded closed or securely locked. Secondly, manual control systems can be locked out with a pass code or key so that pirates cannot change the ship's course, only damage its systems to stop it moving. Thirdly, if pirates did board the vessel, they would have no hostages to protect them and would thus be open targets - as a last resort the vessel could simply be bombed or torpedoed. Lastly, the vessel could be booby-trapped with no fear that a crew member might accidentally be injured.
It is possible to have a broad conversation about what work is for, whether it is purely to provide food, shelter and sanitary facilities, or whether it is to provide people with a sense of purpose, or perhaps it has become a means of determining status in lieu of ceremonial dancing, fighting or willy waving.
Some ways of starting such a conversation are better than others. For example, famous Welshman Bertrand Russell penned an essay titled The Case for a Leisure Economy. A fair gambit. Conversely, just writing 'utterly stupid' just doesn't cut it.
Also, not as sexist as it may seem! Yvette Amos, for example, did an exemplary job for Wales recently.
To those arguing that automation is bad because it puts people out of a job.
It is the idea that we should keep doing work that no longer needs to be done in order to preserve an economic model that is no longer appropriate that is the stupid & short-sighted view.
Similar to a man who will not have a washing machine, dishwasher or vacuum cleaner in the house because having his wife do hard labour is the only way he can justify buying her food & clothing.
>Similar to a man who will not have a washing machine, dishwasher or vacuum cleaner in the house because having his wife do hard labour is the only way he can justify buying her food & clothing.
Those devices are popular because they work a lot better than the manual equivalents and they don't break down that often. If you've ever watched any of the videos of the life of a typical merchant mariner then you'll notice that very little of their activity is concerned with sailing the ship. They're always fixing something or performing preventative maintainance. This is what makes autonomous ships a bit of a nonsense -- the humans sailing that ship are only there to help dock it and carry the can when something goes wrong.
Incidentally, should you find yourself associated with a 'wife' you'll find out pretty quickly that they're actually a form of person, a partner rather than an employee and one that's likely to take a dim view of being regarded as cheap labor. They also tend to have control of the domestic budget.
I’ve been involved in several autonomous shipping projects at Kongsberg for the last 18 months (since I joined the company).
https://www.kongsberg.com/maritime/support/themes/autonomous-shipping/
Inertial navgation is not really usable for long distances when you're on a ship that's constantly rocked by waves, pushed around by winds or even dragged by sea currents.
It works better for submarines in immersion because they only have to deal with underwater currents, which are relatively constant.
What's left for a civilian ship is GPS or automated celestial navigation, both can be unreliable.
Other navigation methods can be used for gross error checking. The magnetic compass can check whether the GPS derived "course made good" is within sensible limits. Inertial navigation can ensure that the GPS is not giving a position that differs by many miles over a few hours, and if necessary systems such as Decca and Loran could be revived for gross error checks as well.
On top of which the operating company can observe real-time satellite tracking of the vessel via its AIS.
If you're spoofing GPS to divert a ship somewhere, I think AIS is just going to report the spoofed position, so the operator won't know anything is wrong. Now if another ship has to divert around the spoofed AIS ship, they may notice the something is amiss if they can't physically see the spoofed ship.
Interesting that Samsung is starting with a difficult route. Tricky and short routes seem better for humans. Long and boring stretches of open ocean seems to be a better use case, even if you needed to board a human crew for the last few miles into harbor.
Maybe that's the point, if it can handle this route, open ocean will be easy.
"I think AIS is just going to report the spoofed position"
Yes, it will. But if that position is overly far from its real position, the satellite receiving the signal will "know" something is wrong because it is not in line-of-sight of the reported position, and this could be flagged.
Two decades I knew a chap who'd been second mate on a container carrier. He'd given this up and gone to work on a nuclear reactor (where I met him) because "it was less dangerous". He told me that, on ships with stopping distances measured in miles, nobody on the bridge was capable of navigating without the onboard technologies. They would "drop the pilot, engage the satnav and put their feet up on the console" (his words).
Given the crowded seas and the reliability of transport automation demonstrated so far ("Windows for warships", Tesla autopilot, 737 MAX), I wonder whether autonomous shipping is the best of ideas. It's a much harder task to stay on course at sea than it is in the air, let alone to follow a road, and preventing collisions requires a lot of attention, forethought and judgement.
The bean counters will love the idea of no salaries for crew, although they might need some emergency standby personnel on board to reboot the computer when it crashes.
>>They would "drop the pilot, engage the satnav and put their feet up on the console" (his words)
Yup. Got a relative who used to, in his words, "get paid for putting his feet on the desk and looking out of the window". He was towing oil rigs across the Pacific at the time.
So I too am not sure what the new bit that Samsung Heavy are offering is - perhaps displacing the need for a pilot?
Frankly, the need for a pilot has been a non starter for the last hundred years at least.
Originally, the practice of hiring a pilot becoming widespread was caused by the Royal Navy in 1700-1800 not really wanting to run aground with an expensive and difficult to replace wooden ship. With no effective accurate charts available the safest way of navigating shallow areas was hiring a person with individual knowledge of that area; a pilot. It was then laid down in the naval general printed instructions that when going into harbour thou shalt hire a pilot as the cost of doing so for a few hours is cheap (even at consultancy rates) compared to buying a new ship.
And if thou break the ship entrusted to your care then the part of your commission that says "Hereof nor you nor any of you may fail as you will answer the contrary at your peril." comes into play. Noticing the relatively fewer accidents this occasioned eventually the insurers at Lloyds of London required commercial outfits to do the same. (via saying that if you ran aground in a harbour where a pilot might have prevented the accident then your insurance was void)
Now this all makes sense, up until every bit of water on the planet was nicely mapped so everybody knows where all the shallow bits with rocks sticking up are. At this point nobody had to rely upon the local knowledge of a pilot; they could get the same from a chart.
As a job, that had pretty much happened by 1900 even with the job being done with people in boats tossing lead weights with depth marked string overboard. With the advent of tide charts, sonar and photos from orbit? I'm not quite sure what relational reason there is for a pilots job to exist anymore.
"I'm not quite sure what relational reason there is for a pilots job to exist anymore."
And yet they do. They're compulsory in most ports. If you think all you need is a chart and a photo to take a 20,000 TEU box boat up Southampton Water I've got a bridge I'd like to sell you.
"Now this all makes sense, up until every bit of water on the planet was nicely mapped so everybody knows where all the shallow bits with rocks sticking up are. At this point nobody had to rely upon the local knowledge of a pilot; they could get the same from a chart."
Yes, but sandbanks are known to shift rather rapidly.
@IGotOut
"Makes far more sense than say a driverless car."
While driverless car would be gimmicky with potential improvements for people, I think automated trucks are onto a winner even if they had certain limitations as to which roads they can use. Automated ships and automated trucks would make for great improvements in shipping worldwide. Longer distances, fewer breaks and consistency would improve everyones lives probably world wide
@Phil O'Sophical
"Isn't that essentially a freight train?"
It used to be. But we have a network of roads with greater reach and the costs of train lines is huge (HS2 for example). But yes it would probably serve as a better version of the freight train (unless the source and destination are already really close to the loading facility).
I know some people who took a US Coast Guard course that covered open water navigation rules. Attendees included fishermen and recreational sailboaters. At the end of the class, the CG officer said something like "listen to me. I just taught you all of the official rules of way, but here's the most important one: bigger wins".
I disagree that "intelligent" autopilots on ships are more difficult to implement than on aircraft, or that the probability of collision is higher. Ships may need tens of miles to stop, but aircraft have an infinite stopping distance! Vessels equipped with AIS (i.e. just about all vessels these days, even including large buoys and most small yachts that venture offshore) can detect a potential collision situation tens of minutes away, and the navigation/autopilot computer can plot a course between moving vessels even in pretty crowded waters well within the time needed to make the appropriate course adjustments, whilst aircraft may have only seconds to detect and react to a collision situation.
The added safety of having humans on ships is debateable. In fact the *illusion* of increased safety may well cause it to be *less* safe. Ships (like airliners) are being controlled by autopilot 99% of the time anyway, and human lookout is more fallible than an automated collision detection system based on radar and AIS - especially so in fog and heavy squalls. When sailing a small yacht (my only personal experience of navigating a vessel at sea), I would far prefer to rely on an AIS plotter or radar proximity alert than my eyeballs. It's very easy to daydream or doze off, or fixate on looking in the wrong direction, or become engrossed in a book, meal preparation or passage planning etc.
Leaving and entering a port and docking is something else entirely, and I doubt that can be reliably automated - but unlike aircraft, a pilot crew can be taken off a ship after clearing port, and put on as it approaches its destination, leaving the ship unmanned during the days or weeks of the en-route portion of its passage.
Watching a Red Funnel car ferry arrive at East Cowes when the water is really active with sailing craft is sometimes fun, and sailing boats have found that they can’t get out of the way quick enough occasionally.
https://www.yachtingmonthly.com/news/investigation-after-red-funnel-ferry-sinks-a-yacht-in-cowes-harbour-67771
But they don't need to follow a road, they have an entire 2D ocean surface to move across so the odds of accidentally hitting something are infinitely less than a car that's restricted to 1D roadways.Sure it is busier near ports, but many ships are towed in/out of ports anyway, or a remote operator could take over direct control for that task.
While humans might get sleepy and not notice a new radar contact during a long night watch, the ship never sleeps so it could steer around anything sitting in its path like a drifting ship with a damaged engine. Anything it is uncertain about can have all the radar/visual/infrared/sonar etc. data relied back to home base for someone to manually direct to change course if necessary.
Back in the day when you were able to go to these strange places that served beer and food.
This was the Cruising Association in Limehouse and a guy was giving a talk about how all ships would be at least semi autonomous soon. He scared me as he had no concept of secure design. Every time I or someone like me pointed out that most computer networking was inherently insecure by design he waffled on about how you would have to patch things regularly. He was also far too excited about how a small camera could give amazing target definition. I suspect that distinguishing between a small boat and a LANBY is going to be a struggle when said software is only expecting the LANBY.
A "crashed" oil tanker is going to bring a new angle to the Bork, Bork, Bork column!
That these behemoths will still need at least a small crew if only to make sure it doesn't suddenly stop working. Freak out over conditions beyond it's comprehension etc etc.
I mean, planes can fly themselves from start to finish with very little input. But you don't see pilots getting the boot anytime soon™ (not to mention the amount of trouble automated drones have been causing).
And that's before you consider things like bits of the scenery changing from time to time like you do at sea (think sandbanks for example) since all maps are only as good as the last time they were drawn up (and even then they can still be wrong).
Crew...
Apart from the routine engineering and maintenance that goes on round the clock/during a voyage, the crew on container ships go round checking the containers at regular intervals like the lashing points and if a refrigerated container, then those too are monitored and attended to if required. What an autonomous ship would give is lessen the workload/and or reduce the crew level for the bridge.
As much as I love the idea, and applaud Samsung for having the balls to actually set this up and try it, I can only think of one thing : what do you do when it goes wrong ?
If it works, fine, congrats, job well done. But if it doesn't, what are the possible consequences ?
Best case scenario you've got a beached ship. Expensive, but no harm done.
Worst case scenario you've got a collision with loss of life. Not only expensive, but a massive lawsuit in the making.
I'm pretty sure I wouldn't want to be the lead developer answering questions in front of that jury.
"well shy of the 200,000-tonne-plus behemoths that roam the oceans. Scaling to handle such craft will take time"
Why will driving bigger ships take more time to learn? Stopping distance and turning circles are a bit worse, but other than that everything is essentially identical. Either your ship can drive where it's supposed to, or it can't. The consequences of screwing up might be worse with a bigger ship, which is presumably why they're testing with smaller ones, but once they're confident the system works it should be possible to plug it into any ship you like and have it work just as well.
"No they don't. Size has nothing to do with it. Sail has right of way over power."
No they don't. It's a bit more complicated than you seem to think. Unpowered boats generally have right of way over small powered boats. Large commerical ships, commercial fishing, and tugs generally have right of way over other boats regardless of whether they're powered. Emergency craft have right of way over everything.
The general rule of thumb is basically just common sense - the craft most able to manouevre is the one that should do so. A sailing dingy doesn't have right of way over a supertanker, because the tanker is simply not capable of steering out the way in time even if it wants to.
It is the ocean. Salt water. Sun & weather exposure. Stuff breaks. I'm not worried about the computerized navigation controls. I'm thinking about the unexpected maintenance problem that happens 1000 miles from any land.
Go watch some documentaries about life on a cargo ship. The daily routine is repeatedly making sure everything works (including things like refrigeration on cargo containers), and then fixing it when it doesn't. There is a certain amount of intelligence needed to recognize a problem before it happens and I don't think automated ships can perform that.
crew often hard to come by .. only if you are not willing to pay a reasonable wage for someone to do a highly technical, unglamourous and dangerous job.
I love these stories, automation is already at sea, in the engine room, cargo handling and the bridge. The main role of ship's crew these days is handling the jobs that it is too expensive to automate when you already have humans on board, hence the lean manning.
We the ship's crew (I am a ship's electrical engineer) are here to do routine at sea maintenance, make decisions, and most of all try and save the ship in an emergency (because no one wants to be in a lifeboat and the owner/customers would like to see the cargo).
No one has so far mentioned the timescale in the article. They are hoping to trial this year and start selling it as a commercial offering next year.
Hey Samsung! This isn't a fucking new model of mobile phone. This is a system that could potentially be controlling fucking big ships that need to navigate ports, not just open seas or even channels between islands. I'd expect it to take a lot longer than a year from trying it out to properly analyse the resulting data, let alone patching, tweaking, updating, fixing, trying again then re-iterating till you get at least nearly right.
Icon for what might happen if you rush into this and install the system on an LPG/LNG tanker before it's ready.
Selling a system does not mean that it will be instantly installed and used autonomously. The buyer will no doubt have to wait for delivery & installation (maybe a year or two), then conduct their own trials - probably with many runs with a safety crew on board, and only after many successful passages will they actually operate a completely unmanned ship.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
