Autonomous Boats Could Service Some Cities, Reducing Road Traffic | Hacker News Reader

Autonomous Boats Could Service Some Cities, Reducing Road Traffic

cacm.acm.org

34 pointsrbanffy37 comments

This article reads like a freshman essay. Autonomous boats! Self assembly! 3D printing!

Ignores the big issues: * For large boats, fuel is expensive. The person operating the boat is a tiny fraction of the cost. * Maybe electricity is the idea? The article doesn't say. * Shifting city garbage to the night? Collecting aggregate garbage isn't really the problem. The problem is getting it from the alley to the truck. No boat is going to help here, unless you're in Venice.

This article doesn't make a lot of sense to me. Even the model doesn't accurately represent the dreams of the article; it clearly is not a 2x4m hull.

I have used NYC ferry service, and it is awesome. It is a resounding success with people that use it, but it costs the city money. (The fare is the same as subway $2.75 but the true cost is $5.60+ per trip).

https://www.ferry.nyc/routes-and-schedules/

The other problems is that it doesn't run often enough, and in good days it can get overcrowded, but in bad weather days it is underused.

Having a fleet of autonomous boats will allow to:

1. Run the service more often, (smaller but more frequent boats)

2. Have more routes/trips

3. Scalability. Have more boats ready for peak demand. (it is harder to scale when you have hire and train humans, but if it is all autonomous, that it is simply of an issue just buying more hardware and deploying them on demand).

They might eventually be able to cut costs a little by automating the job of the helmsman (coxswain) in confined waterways where there are few navigational hazards. But automated ferries will still need trained seamen on board for safety to deal with the situation in case of a man overboard, flooding, fire, or power loss. A boat isn't like a car where it can just pull over on the side of the road and the passengers get out.

The USN though they could skimp on training on ship operations.

Oh lets see that's not working so well as quite a number of captains and senior staff are getting (rightly) crucified at courts martial.

Don't forget more but smaller boats, since each boat would no longer have the same overhead.

Best $2.75 you can spend in the NYC.

Depending on route, it can be a 30 minute boatride alongside the world's most iconic skyline. Incredible value in a city where so much feels painfully overpriced.

And the Staten Island ferry, with its spectacular views of lower Manhattan, One World Trade, and the Statue of Liberty is free!

It’s a fantastic way for a tourist to spend an hour at no cost.

I wonder what percent of the fare is consumed by automatable employees? Even if you could replace the captain, you would still need other staff on board to help people and deal with any kind of emergency.

> The fare is the same as subway $2.75 but the true cost is $5.60+ per trip

What is the true cost of the subway trip?

Also, I'm surprised a boat is more expensive. Generally speaking, shipping things by water costs far less; that's why access to navigable waterways is an essential economic resource. I could imagine boats in NYC would have all sorts of overhead, but does anyone know what it is?

Nope. Never. Not going to be a thing. Have a look at any busy harbour. Look at all the various ships, aircraft, and even animals moving about. Ships need humans onboard. They need the adaptability to handle strange situations in strange circumstances. They also need to anticipate how the other human-operated ships will react, something AI is horrible at atm. And at night? It doesn't get any easier at night.

Harbours are the hardest part for ships. Built me an autopilot that can reliably handle ocean crossings (not course holding, actual safe ship operation) first before telling me that it can handle a busy harbour.

> Harbours are the hardest part for ships

I know almost nothing about it, but don't ships have to hire local pilots to navigate harbors, due to the local knowledge and complexity?

Yes:

https://www.nytimes.com/2011/06/16/nyregion/an-apprenticeshi...

That dude is linehandling with a tie and white shirt. HAHAHA. Fake new site so it doesn't surprise me. I deal with the new sailors who come out of these maritime institutes. There is no substitute for OJT.

Can you imagine getting the tie pulled into the wildcat. Youch. That has to be a OSHA no-no. Dopey picture. Like $70K fine in my world.

Actual article (this one is a teaser): https://news.mit.edu/2018/fleet-autonomous-boats-service-cit...

why do these boats need to be autonomous?

Non-autonomous boats already service some cities (e.g. https://www.reuters.com/article/us-dhl-amsterdam-boat/in-ams...)

MIT cooperates with the Amsterdam Institute for Advanced Metropolitan Solutions (http://news.mit.edu/2018/fleet-autonomous-boats-service-citi...), so they likely know that doesn’t need researching.

For that matter, the Boston area has all sorts of ferries and water taxis including at least one that services the airport. There's no routine local delivery of goods by water as far as I know but there's routine commuter traffic.

In San Francisco the typical ferry requires at least 3 crew members. There's a captain, two deckhands and sometimes a coast guard officer (I'm assuming to prevent hijackings which an autonomous system wouldn't be liable to).

A few years ago when BART workers went on strike the ferry service had to call in all available workers. The people they brought in were clearly amateur, they needed a lot of supervision and seemed to struggle with assisting with the docking/undocking procedures. Boating requires a lot of labor, and it seems like there isn't much of it that's trained and ready to work.

Why doesn't SF have a better ferry system? Almost everything in the bay area is within 5min of... the bay. Seems like a no-brainer. What's the hold up?

Outside of San Francisco, a lot of the bay shoreline is not much more than acres of foot-deep mud and salt flats.

http://www.charts.noaa.gov/OnLineViewer/18651.shtml

Sierra Point, Coyote Point, Redwood City are the only current viable ports along the south. Creating new ports for a ferry system would involve a lot of dredging in addition to the infrastructure construction, which is not only costly but severely impacts the already damaged wetlands that line the bay.

Hovercraft or fan boats. Something with skis that can go on mud.

I wonder if hovercraft might be an option that wouldn't require the dredging. I.e. https://www.hovertravel.co.uk/

EDIT: I guess it's been considered before in 2011 [1] But it appears that the hovercraft themselves were more expensive and carried half as many people. But compared to the cost of dredging for ferries, maybe it's still cheaper?

1. https://www.mercurynews.com/2011/11/01/east-bay-looks-to-hov...

> But compared to the cost of dredging for ferries, maybe it's still cheaper?

I think the point is that it doesn't matter, if it's simply not cheap enough.

The shallowness of the estuarial coastline significantly drives up the cost (be it with dredging or more exotic technologies like hovercraft), and that's the (short) answer to why we don't have more ferry service.

Why wouldn't autonomous ferries be liable to be hijacked?

This is really nice and all but I'm pretty sure shortage of workers is not the reason this doesn't already happen. But if this helps to raise money / grants, I'm for it.

It may not be a shortage of workers, but that hiring the required workers makes the boat too costly to operate at the price point they want.

That said, given the number of variables you have to consider when operating watercraft it seems like a challenging problem to completely automate.

I could see a company like Disney using autonomous boats for WDW as it’s all a pretty confined area. If you haven’t been there, they use a lot of boats to ferry people around. (Actually think they already are using some autonomy for their show rafts but not 100% sure)

First time in a while an HN headline made me laugh.

I’ve lived in the Finger Lakes of Upstate New York most of my life. I’ve sailed, done jet skis, onboard motor, offboard, name it. Loads of boating experience all around.

And yet every single time I dock it’s always a struggle to time and execute it exactly right, even with fair weather, even with someone else to jump off and tie the boat up to the dock.

Autonomous ferries are the dumbest idea I have read on this site in a long time. The complexity is light-years beyond the trials and tribulations of Tesla’s “Autopilot.” People could drown, or at least keep themselves afloat while the water ruins whatever electronics they brought aboard.

I would sooner trust an autonomous car than an autonomous ferry, and that’s after reading about those Uber/Tesla crashes.

Hello from Boston! Come see how gridlocked all modes of transportation are during the day and suddenly this idea will shift from the "that's impossible" category to "how can we make this possible" category. We're considering gondolas for public transit, for example. All options should be on the table.

Have you ever tried to fly a quadcopter drone manually? It's really hard; I could barely get one to hover, after hours of practice. Yet AI quadcopters that can fly in ornate formations and execute complex maneuvers are well known [1]. Don't discount what AI can or can't do just because it's hard for humans.

Other advantages in ferries' favor; unlike most streets, the docks for autonomous ferries could be designed around the AI's capabilities. Also, ferries don't travel at 70 MPH. :)

[1] https://www.youtube.com/watch?v=YQIMGV5vtd4 (6-year-old video!)

Things not in ferries' favor: wind, water currents, depth, other shipping which is often not very maneuverable and yet moving fast, no lanes, complex rules about giving way to other boats, and no ability to fail safe (just turning everything off is dangerous, possibly deadly).

Ferries crash real good, too (read to the end, about the second crash):

https://en.wikipedia.org/wiki/2003_Staten_Island_Ferry_crash

Did the ferry's crash because they were autonomous?

No quite the opposite. Human error.

I think that Autonomous ferries are a better idea than Autonomous Tractor-Trailer trucks going down the crowded freeway.

For machine vision the fact that there are few obstructions on the water make this a killer testbed. Have you seen the quality of detail on marine charts?

I work in a shipyard and understand docking is difficult. It is an easy engineering challenge though. If we can do it on the space station then it is achievable on the nations waterways.

“Some cities” isn’t “every city”, and this article isn’t talking about ferries or about big harbors. I can see this having more advantages in a medieval city with narrow streeets than in a city designed for cars.

I imagine that leaving the mooring to human operators (either those waiting to (un)load the ship or remotely) also could be an option.

More importantly, we’re talking research. This need not lead to a economically viable product in decades.

Personally I would freak out. It would be just like being in an airplane completely on Autopilot with no pilots even for emergency. Except that this is on water, and not air.

It would certainly be a novel way to kidnap a bunch of people over the Internet.

This has the same problem as Musk's pneumatic tubes. When it fails, people are trapped. One mistake and you have a boat full of dead people.

Reply on news.ycombinator.com