Live data from Hacker News

Autonomous Ships

bloomberg.com

41–50 of 74 posts

Re: Autonomous Ships

#41

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?

Look into flexport.com they already do this.

Re: Autonomous Ships

#42

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 als…

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).

Re: Autonomous Ships

#43

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 als…

>With no controls on the ship, pirates wouldn't be able to take control.

So they'll just figure out how to disable the ship. And then they have a big drifting fortress to squat on/ransom/move cargo off of.

Re: Autonomous Ships

#45

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?

isn't Flexport doing that? or am i mistaken?

Re: Autonomous Ships

#46
post #32
post #14

Earlier quoted context omitted.

"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 li…

While I agree with your analysis, one counter-argument I can think of: Is the ship always using that much fuel? I know a lot of ships spend days waiting to enter the panama canal, or days in port. Crew is certainly paid for those days as well. What percentage of the time is the ship fully underway and using fuel? That 44% figure might be more reasonable if you're only burning fuel 1/3 of the time.

I did think of that, and also that you may need other shifts on retainer, and the normal food/insurance/fees/overhead of employing people across borders. Still though, for the triple E the expected fuel cost is 30x the expected crew cost.

The stretches to explain that gap would be ludicrous eg 3 shifts, each costing the company 3x more than their actual salary, and the engine being turned off 70% of the time. Even still that requires you to disregard depreciation, which for the triple E is at least $17,000 a day.

Re: Autonomous Ships

#47

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 als…

>With no controls on the ship, pirates wouldn't be able to take control. So they'll just figure out how to disable the ship. And then they have a big drifting fortress to squat on/ransom/move cargo off of.

Taking hostages is the only reason why pirates can (sometimes) ransom the ship instead of getting cleaned off by PMCs or even foreign militaries patrolling the waters.

A ship without hostages means that a pirate has a good chance of not getting home ever again - the patrols aren't allowed (rightly) to stop a "fisherman" boat in these waters before it's done anything, but once they are on board and making a ransom note, they'd be lucky if they survive to be arrested.

Re: Autonomous Ships

#48
post #39

Earlier quoted context omitted.

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.

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.

How does it work currently? So much redundancy that there's never an unfixable failure that would stop the ship?

Re: Autonomous Ships

#49

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?

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
Post reply on HN