8 ms·

Autonomous Ships

37 points by protomyth 9y ago

LeonM 9y ago
As far as my knowledge goes about the shipping industry: ships are already 90% automated (since they simply go straight forward for days/weeks). There is not much to do for a captain/officer whilst in the middle of an ocean. AFAIK most of the humans on board of a vessel are there to perform maintenance (most of which can be done during transit). Yes, we can automate the last 10% to make the ship sail itself, but we'd still need humans on board to perform maintenance.
- keebEz 9y ago
  Maybe fully autonomous ships should have maintenance performing robots then?
  - zardo 9y ago
    That is a way harder problem than automated sailing. Also a way harder problem than just designing in longer service intervals.
  - noir_lord 9y ago
    That or have a small service vessel that sails with a school of autonomous ships handling maintenance.
  - ghaff 9y ago
    Honest question. Have you ever been in the engineering spaces in a ship/offshore rig/etc.? Do you have any idea of the diverse type of work a ship's engineer and his various reports do? I don't mean to pick on you too much but the idea that you can have little Hueys, Deweys, and Louies tooling around a container ship doing whatever maintenance is required is ludicrous. And the other suggestions about having crews that you transport from ship to ship isn't quite as unrealistic but is needlessly complicated in the way that software often is. Among other things--have the people suggesting this sort of thing ever been in the North Atlantic in bad weather?
thomasthomas 9y ago
By one consultant's estimate, moreover, carrying sailors accounts for 44 percent of a ship's costs. I find this hard to believe even if autonomous ships can be done without a bridge. The Emma Maersk has 15,000 TEU and a crew of 13 http://www.emma-maersk.com/specification/ http://www.emma-maersk.com/specification/
- kingbirdy 9y ago
  A crew that has to be paid, fed, insured, provided habitable accommodations on the ship including sleep, cooking, and personal hygiene, and provided benefits. I'm sure that's not cheap (though perhaps not 44%, either).
- hwillis 9y ago
  "Crew costs of $3,299 a day account for about 44 percent of total operating expenses for a large container ship, according to Moore Stephens LLP, an industry accountant and consultant."[1] The Maersk triple E's have a crew of 13 and consume ~136 tonnes/day of bunker fuel[2] which costs currently ~$315/tonne[3], for a daily fuel cost of $42,840. So right off the bat that number is very suspect. Even for other ships like the CSCL Globe with a crew of 31, that's a 12x difference. I don't understand how that estimate can possibly be true. Ship captains and engineers make ~80k annual and crewmen make under 40k. For a crew of 20-30, $3,299 would be basically bang on for a daily crew cost but there's no fucking way fuel and depreciation only cost $4,200 a day. Even a 4000 TEU ship (tiny) at 17 knots (snail) burns $15,000+ a day.[4] [1] https://www.bloomberg.com/news/articles/2014-02-25/rolls-royce-drone-ships-challenge-375-billion-industry-freight https://www.bloomberg.com/news/articles/2014-02-25/rolls-roy... [2] http://www.scdigest.com/ontarget/13-09-12-1.php?cid=7401 http://www.scdigest.com/ontarget/13-09-12-1.php?cid=7401 [3] http://www.tsacarriers.org/calc_bunker.html http://www.tsacarriers.org/calc_bunker.html [4] https://people.hofstra.edu/geotrans/eng/ch8en/conc8en/fuel_consumption_containerships.html https://people.hofstra.edu/geotrans/eng/ch8en/conc8en/fuel_c...
  - electic 9y ago
    What about food? What about living supplies? Also, dock living services? And other dock related expenses for crew? Don't forget, all this has to be insured on sea and on land.
JohnLeTigre 9y ago
As guess this would qualify as great news for sea-faring pirates.
- keyme 9y ago
  Hacking shipping route info, then sending pirate boarding parties would be pretty cyberpunk (and too easy for these drone ships to actually catch on).
- remarkEon 9y ago
  But what if we just reduced the crew to a few security personnel? Don't want to suggest automated security systems...too terminator for me at this point.
- marcosdumay 9y ago
  Or maybe not. There's also nobody to make hostage or force changing a course.
- eeks 9y ago
  A fully weaponized autonomous ship would be pretty badass.
  - tonyedgecombe 9y ago
    https://www.theverge.com/2016/4/8/11391840/us-navy-autonomous-ship-sea-hunter-christened https://www.theverge.com/2016/4/8/11391840/us-navy-autonomou...
  - dredmorbius 9y ago
    Until commandeered electronically within your harbour.
- bm1362 9y ago
  Most pirates aren't actually after cargo I'd assume. They're just taking hostages and ransoming them.
  - 9y ago
noir_lord 9y ago
I would love to work on this kind of software. It's exactly the type of things I find interesting, it's a shame I didn't make different career choices when I was younger but it is what it is I guess.
- askafriend 9y ago
  What makes you think you can't work on this type of software?
  - noir_lord 9y ago
    I'm 37, I have few academic qualifications so to transition would require 4-5 years more education and then I'd be a 42-43 year old competing against guys coming straight out of university with a background in desktop and web application development. Practicalities really.
orbitingpluto 9y ago
"The U.S. Coast Guard has estimated that human error accounts for up to 96 percent of all marine casualties." But if we somehow manage to account for 90% of all eventualities when programming, how are we going to be safer?
melling 9y ago
Prior HN coverage: https://news.ycombinator.com/item?id=14315036 https://news.ycombinator.com/item?id=14315036
otto_ortega 9y ago
I'm interested in taking a shot creating a SaaS business related to the freight forwarders / containers shipping logistic or similar. Does anybody here works on that sector and is willing to share some of the troubles they perceive on the industry that can be solved with software / machine learning?
- Allvitende 9y ago
  Look into flexport.com they already do this.
- forgotmysn 9y ago
  isn't Flexport doing that? or am i mistaken?
- otto_ortega 9y ago
  it seems that Flexport is a freight forwarder on its own, I was thinking of something more in the line of solving problems logistics operators have without forcing them to change providers
chriskanan 9y ago
I spent two years working on autonomous ships at NASA JPL, primary on developing deep learning algorithms for ship detection and classification to obey maritime traffic rules (COLREGS), avoid obstacles, help with decision making, etc. Regarding piracy, autonomy would cut down on costs a lot. Based on my understanding, besides cutting down on life-support costs and having extra room for more cargo, the ships would also be able to eliminate the cost of the armed guards that many have kept on their ships since the rise of piracy. That's one of the main reasons why we don't hear about piracy anymore [0]. With no controls on the ship, pirates wouldn't be able to take control. Note also that Rolls Royce is looking into drone ships (rather than full autonomy): https://www.rolls-royce.com/media/press-releases/yr-2016/pr-2016-03-22-rr-reveals-future-shore-control-centre.aspx https://www.rolls-royce.com/media/press-releases/yr-2016/pr-... [0] http://www.marineinsight.com/marine-piracy-marine/18-anti-piracy-weapons-for-ships-to-fight-pirates/ http://www.marineinsight.com/marine-piracy-marine/18-anti-pi...
- oconnore 9y ago
  Drone ships sound interesting. The shipping company can definitely afford having a person remotely check in on millions of dollars of cargo every ~15 minutes or so, and so full autonomy seems like the 20 in 80/20 (maybe more like 95/5).
- jdavis703 9y ago
  What would stop heists on the high seas? For example, I could imagine someone breaking into a container with "valuable" goods (remember a lot of these pirates come from impoverished backgrounds, so being able to flip cargo still net a nice profit).
  - blacksmith_tb 9y ago
    Possible, but unloading cargo would be more involved. Also, without access to electronic manifests, I am not sure there's any easy way to know what's in a container, not to mention many of them would be stacked in such a way as to inaccessible. All of which is a lot harder than ransoming a crew. I suppose it might be possible for pirates just to threaten to sink the ship.
nradov 9y ago
I don't get it. What will they do when an autonomous ship loses power or steering in the middle of the ocean outside of helicopter range? It can't just pull over to the side of the road and stop. Ships require constant maintenance to not fall apart. That maintenance can't be automated with anything like current technology. If the crew doesn't do maintenance en route then it will have to be done at the dock, thus taking the ship out of service and costing more money.
- wlesieutre 9y ago
  Why's it need to pull over? It's in the middle of the ocean. Unless you're a spaceship it's hard to get any more out of the way. Presumably you send out a repair crew by boat.
  - nradov 9y ago
    No that's not how it works. Disabled ships don't just sit there. The wind and currents eventually push them into rocks. Disabled ships also present a hazard to other vessels. Repair boats don't move fast enough to get there in time, especially in rough weather.
    - wlesieutre 9y ago
      How does it work currently? So much redundancy that there's never an unfixable failure that would stop the ship?
      - nradov 9y ago
        Unfixable failures are very rare in critical systems. Usually the crew is able to catch incipient failures and take preventative action, or else repair the failure at least well enough to get close to a port. Ships engineers are good at jury rigging temporary solutions. There are a few incidents every year where large merchant vessels lose power and drift onto rocks, or break apart and sink in a storm.
        9y ago
otempomores 9y ago
Lets call the sailors mechanics and the ships frail tech in harsh environment. Suddenly not a good idea.
Casseres 9y ago
Former Marine Engineer here, I know there's at least one other HN users that's an ME. You might be able to automate navigation, but never maintenance. Something small that doesn't get fixed underway could cause a lot more damage than whatever crew costs would have been saved. Things break and leak all the time, even on new ships. Automatic valves get stuck and things get clogged. I'm an air traffic controller now, and my break is over, but I'll try to expand more later. Edit: I have a few minutes. A lot of people are talking about pirates only boarding ships to hold the crews hostage, and not to steal cargo. How about this: a pirate demands a crypto currency payment within X number of hours, or they sink the ship. It's very easy to sink a ship once you're inside.
- MurMan 9y ago
  I will agree with you that steam or turbine power plants require continuous maintenance. I've got experience, too. :-) But, read the article. The ship they're talking about is a short-haul transport with battery-electric propulsion. Perfect scenario to pioneer this concept. If it works, I'm betting that it would be possible to create a propulsion plant that will survive longer hauls without a crew.
- hwillis 9y ago
  You could fly a helicopter out to any ship that isn't crossing the pacific or atlantic, but that would be $100,000-$300,000 per incident and it would very quickly not be worth it. There isn't much in-between for reducing crew since a couple people alone on a boat together will probably be homicidal in short order. I'm curious, what breaks on a ship? I have no experience with anything over a few dozen feet. I imagine hydraulic systems could break and lead to Consequences, but I can't imagine there are many big systems that can be fixed at sea, even with a crane onboard. What kind of things get fixed at sea?
  - nradov 9y ago
    You might as well ask what doesn't break on a ship. Salt water corrosion, thermal stress, vibration, residue accumulation, and hull flexing due to large swells are really hard on equipment. Here's a good case study on the type of repairs that a merchant ship crew might have to perform in the middle of the ocean. http://www.brighthubengineering.com/marine-engines-machinery/91635-breakdown-and-repair-of-ships-main-diesel-engine-at-sea/ http://www.brighthubengineering.com/marine-engines-machinery... Ocean going merchant ships spend much of their time outside of practical helicopter range. And even when close to shore, winching a repair technician down to a disabled autonomous vessel that can't even keep its bow into the weather would be a dangerous affair.

37 points by protomyth 9y ago

As far as my knowledge goes about the shipping industry: ships are already 90% automated (since they simply go straight forward for days/weeks). There is not much to do for a captain/officer whilst in the middle of an ocean. AFAIK most of the humans on board of a vessel are there to perform maintenance (most of which can be done during transit). Yes, we can automate the last 10% to make the ship sail itself, but we'd still need humans on board to perform maintenance.

keebEz 9y ago
Maybe fully autonomous ships should have maintenance performing robots then?
- zardo 9y ago
  That is a way harder problem than automated sailing. Also a way harder problem than just designing in longer service intervals.
- noir_lord 9y ago
  That or have a small service vessel that sails with a school of autonomous ships handling maintenance.
- ghaff 9y ago
  Honest question. Have you ever been in the engineering spaces in a ship/offshore rig/etc.? Do you have any idea of the diverse type of work a ship's engineer and his various reports do? I don't mean to pick on you too much but the idea that you can have little Hueys, Deweys, and Louies tooling around a container ship doing whatever maintenance is required is ludicrous. And the other suggestions about having crews that you transport from ship to ship isn't quite as unrealistic but is needlessly complicated in the way that software often is. Among other things--have the people suggesting this sort of thing ever been in the North Atlantic in bad weather?

thomasthomas 9y ago

By one consultant's estimate, moreover, carrying sailors accounts for 44 percent of a ship's costs. I find this hard to believe even if autonomous ships can be done without a bridge. The Emma Maersk has 15,000 TEU and a crew of 13 http://www.emma-maersk.com/specification/ http://www.emma-maersk.com/specification/

kingbirdy 9y ago
A crew that has to be paid, fed, insured, provided habitable accommodations on the ship including sleep, cooking, and personal hygiene, and provided benefits. I'm sure that's not cheap (though perhaps not 44%, either).
hwillis 9y ago
"Crew costs of $3,299 a day account for about 44 percent of total operating expenses for a large container ship, according to Moore Stephens LLP, an industry accountant and consultant."[1] The Maersk triple E's have a crew of 13 and consume ~136 tonnes/day of bunker fuel[2] which costs currently ~$315/tonne[3], for a daily fuel cost of $42,840. So right off the bat that number is very suspect. Even for other ships like the CSCL Globe with a crew of 31, that's a 12x difference. I don't understand how that estimate can possibly be true. Ship captains and engineers make ~80k annual and crewmen make under 40k. For a crew of 20-30, $3,299 would be basically bang on for a daily crew cost but there's no fucking way fuel and depreciation only cost $4,200 a day. Even a 4000 TEU ship (tiny) at 17 knots (snail) burns $15,000+ a day.[4] [1] https://www.bloomberg.com/news/articles/2014-02-25/rolls-royce-drone-ships-challenge-375-billion-industry-freight https://www.bloomberg.com/news/articles/2014-02-25/rolls-roy... [2] http://www.scdigest.com/ontarget/13-09-12-1.php?cid=7401 http://www.scdigest.com/ontarget/13-09-12-1.php?cid=7401 [3] http://www.tsacarriers.org/calc_bunker.html http://www.tsacarriers.org/calc_bunker.html [4] https://people.hofstra.edu/geotrans/eng/ch8en/conc8en/fuel_consumption_containerships.html https://people.hofstra.edu/geotrans/eng/ch8en/conc8en/fuel_c...
- electic 9y ago
  What about food? What about living supplies? Also, dock living services? And other dock related expenses for crew? Don't forget, all this has to be insured on sea and on land.

JohnLeTigre 9y ago

As guess this would qualify as great news for sea-faring pirates.

keyme 9y ago
Hacking shipping route info, then sending pirate boarding parties would be pretty cyberpunk (and too easy for these drone ships to actually catch on).
remarkEon 9y ago
But what if we just reduced the crew to a few security personnel? Don't want to suggest automated security systems...too terminator for me at this point.
marcosdumay 9y ago
Or maybe not. There's also nobody to make hostage or force changing a course.
eeks 9y ago
A fully weaponized autonomous ship would be pretty badass.
- tonyedgecombe 9y ago
  https://www.theverge.com/2016/4/8/11391840/us-navy-autonomous-ship-sea-hunter-christened https://www.theverge.com/2016/4/8/11391840/us-navy-autonomou...
- dredmorbius 9y ago
  Until commandeered electronically within your harbour.
bm1362 9y ago
Most pirates aren't actually after cargo I'd assume. They're just taking hostages and ransoming them.
- 9y ago

noir_lord 9y ago

I would love to work on this kind of software. It's exactly the type of things I find interesting, it's a shame I didn't make different career choices when I was younger but it is what it is I guess.

askafriend 9y ago
What makes you think you can't work on this type of software?
- noir_lord 9y ago
  I'm 37, I have few academic qualifications so to transition would require 4-5 years more education and then I'd be a 42-43 year old competing against guys coming straight out of university with a background in desktop and web application development. Practicalities really.

orbitingpluto 9y ago

"The U.S. Coast Guard has estimated that human error accounts for up to 96 percent of all marine casualties." But if we somehow manage to account for 90% of all eventualities when programming, how are we going to be safer?

melling 9y ago

Prior HN coverage: https://news.ycombinator.com/item?id=14315036 https://news.ycombinator.com/item?id=14315036

otto_ortega 9y ago

I'm interested in taking a shot creating a SaaS business related to the freight forwarders / containers shipping logistic or similar. Does anybody here works on that sector and is willing to share some of the troubles they perceive on the industry that can be solved with software / machine learning?

Allvitende 9y ago
Look into flexport.com they already do this.
forgotmysn 9y ago
isn't Flexport doing that? or am i mistaken?
otto_ortega 9y ago
it seems that Flexport is a freight forwarder on its own, I was thinking of something more in the line of solving problems logistics operators have without forcing them to change providers

chriskanan 9y ago

I spent two years working on autonomous ships at NASA JPL, primary on developing deep learning algorithms for ship detection and classification to obey maritime traffic rules (COLREGS), avoid obstacles, help with decision making, etc. Regarding piracy, autonomy would cut down on costs a lot. Based on my understanding, besides cutting down on life-support costs and having extra room for more cargo, the ships would also be able to eliminate the cost of the armed guards that many have kept on their ships since the rise of piracy. That's one of the main reasons why we don't hear about piracy anymore [0]. With no controls on the ship, pirates wouldn't be able to take control. Note also that Rolls Royce is looking into drone ships (rather than full autonomy): https://www.rolls-royce.com/media/press-releases/yr-2016/pr-2016-03-22-rr-reveals-future-shore-control-centre.aspx https://www.rolls-royce.com/media/press-releases/yr-2016/pr-... [0] http://www.marineinsight.com/marine-piracy-marine/18-anti-piracy-weapons-for-ships-to-fight-pirates/ http://www.marineinsight.com/marine-piracy-marine/18-anti-pi...

oconnore 9y ago
Drone ships sound interesting. The shipping company can definitely afford having a person remotely check in on millions of dollars of cargo every ~15 minutes or so, and so full autonomy seems like the 20 in 80/20 (maybe more like 95/5).
jdavis703 9y ago
What would stop heists on the high seas? For example, I could imagine someone breaking into a container with "valuable" goods (remember a lot of these pirates come from impoverished backgrounds, so being able to flip cargo still net a nice profit).
- blacksmith_tb 9y ago
  Possible, but unloading cargo would be more involved. Also, without access to electronic manifests, I am not sure there's any easy way to know what's in a container, not to mention many of them would be stacked in such a way as to inaccessible. All of which is a lot harder than ransoming a crew. I suppose it might be possible for pirates just to threaten to sink the ship.

nradov 9y ago

I don't get it. What will they do when an autonomous ship loses power or steering in the middle of the ocean outside of helicopter range? It can't just pull over to the side of the road and stop. Ships require constant maintenance to not fall apart. That maintenance can't be automated with anything like current technology. If the crew doesn't do maintenance en route then it will have to be done at the dock, thus taking the ship out of service and costing more money.

wlesieutre 9y ago
Why's it need to pull over? It's in the middle of the ocean. Unless you're a spaceship it's hard to get any more out of the way. Presumably you send out a repair crew by boat.
- nradov 9y ago
  No that's not how it works. Disabled ships don't just sit there. The wind and currents eventually push them into rocks. Disabled ships also present a hazard to other vessels. Repair boats don't move fast enough to get there in time, especially in rough weather.
  - wlesieutre 9y ago
    How does it work currently? So much redundancy that there's never an unfixable failure that would stop the ship?
    - nradov 9y ago
      Unfixable failures are very rare in critical systems. Usually the crew is able to catch incipient failures and take preventative action, or else repair the failure at least well enough to get close to a port. Ships engineers are good at jury rigging temporary solutions. There are a few incidents every year where large merchant vessels lose power and drift onto rocks, or break apart and sink in a storm.
      - 9y ago

otempomores 9y ago

Lets call the sailors mechanics and the ships frail tech in harsh environment. Suddenly not a good idea.

Casseres 9y ago

Former Marine Engineer here, I know there's at least one other HN users that's an ME. You might be able to automate navigation, but never maintenance. Something small that doesn't get fixed underway could cause a lot more damage than whatever crew costs would have been saved. Things break and leak all the time, even on new ships. Automatic valves get stuck and things get clogged. I'm an air traffic controller now, and my break is over, but I'll try to expand more later. Edit: I have a few minutes. A lot of people are talking about pirates only boarding ships to hold the crews hostage, and not to steal cargo. How about this: a pirate demands a crypto currency payment within X number of hours, or they sink the ship. It's very easy to sink a ship once you're inside.

MurMan 9y ago
I will agree with you that steam or turbine power plants require continuous maintenance. I've got experience, too. :-) But, read the article. The ship they're talking about is a short-haul transport with battery-electric propulsion. Perfect scenario to pioneer this concept. If it works, I'm betting that it would be possible to create a propulsion plant that will survive longer hauls without a crew.
hwillis 9y ago
You could fly a helicopter out to any ship that isn't crossing the pacific or atlantic, but that would be $100,000-$300,000 per incident and it would very quickly not be worth it. There isn't much in-between for reducing crew since a couple people alone on a boat together will probably be homicidal in short order. I'm curious, what breaks on a ship? I have no experience with anything over a few dozen feet. I imagine hydraulic systems could break and lead to Consequences, but I can't imagine there are many big systems that can be fixed at sea, even with a crane onboard. What kind of things get fixed at sea?
- nradov 9y ago
  You might as well ask what doesn't break on a ship. Salt water corrosion, thermal stress, vibration, residue accumulation, and hull flexing due to large swells are really hard on equipment. Here's a good case study on the type of repairs that a merchant ship crew might have to perform in the middle of the ocean. http://www.brighthubengineering.com/marine-engines-machinery/91635-breakdown-and-repair-of-ships-main-diesel-engine-at-sea/ http://www.brighthubengineering.com/marine-engines-machinery... Ocean going merchant ships spend much of their time outside of practical helicopter range. And even when close to shore, winching a repair technician down to a disabled autonomous vessel that can't even keep its bow into the weather would be a dangerous affair.