Comment by the_duke | Hacker News Reader

Comment by the_duke | Hacker News Reader

There are a lot of companies working on "flying taxis".

All of those seem to have a large footprint with lots of rotors, meaning they will need dedicated landing/take off zones.

This seems to limit their usability a lot, turning them more into a short range point to point helicopter service.

Not to mention the noise pollution. Those things are all loud.

Combine that with limited speed and range, and I just don't see the concept taking off in a big way.

Some other competitors:

* Volocopter: https://www.volocopter.com https://www.youtube.com/watch?v=OazFiIhwAEs

* Hyunday S-A1: https://www.youtube.com/watch?v=L6K7GAG1Aas

* Bell Nexus: https://www.bellflight.com/products/bell-nexus

* Ehang: https://www.ehang.com/ehang184/index https://www.youtube.com/watch?v=d66MoI4GdFs

* Kitty Hawk HeavySide (Larry Page pet project): https://kittyhawk.aero https://www.youtube.com/watch?v=q7mc3C19kE4

* Lilium (branded as jet): https://lilium.com https://www.youtube.com/watch?v=8qotuu8JjQM

Rotor size matters a lot for noise and how far it carries. Compared to a helicopter these things are going to be a lot less noisy. Also there's a difference between landing/taking off and cruise. Several of these designs have wings and a vtol configuration. There's no shortage of roof tops, open areas, etc. that could be converted to land on. As soon as viable products hit the market they'll be very popular. Finally, cities are noisy places and it will be hard to hear them over the background noise.

The reason this will take off (physically and economically) is cost. If you can travel 40-50m quickly for about the energy cost of a cup of coffee, it's going to be a game changer in big cities. I can take an Uber across town in Berlin but it will cost me around 20-30 euros and take 50 minutes (worse in rush hour) and is not that competitive with public transport. The same journey with a flying taxi could be done in 5-10 minutes (just like with a helicopter) the difference will be vastly lower cost. The main cost will be the pilot who can now do multiple journeys per hour and at least short term still charge a premium.

> travel 40-50m quickly for about the energy cost of a cup of coffee

By "energy cost" do you mean something like "the price of a cup of coffee spent on electricity"? That seems surprisingly efficient but maybe here (Europe) the coffee is cheaper and the electricity dearer - what's the figure in kWh?

Around 20-40 kwh or so for a single journey assuming a maximum capacity of 100-150kwh and assuming it would be rare to use more than 20-30% of it on a single journey or fly the maximum range. At 5-10 cents per kwh (i.e. 20-40kwh/$) we're not even talking some fancy barista doing their thing. Grid prices can be higher of course but I'm assuming a bulk consumer of energy would invest in turning this in a fixed cost by e.g. installing solar panels & batteries to lower the cost.

By the time these things start flying at scale, the coffee will be way more expensive than the energy cost. That's why most of these companies are investing in autonomous flight as well because short term that is going to be the cost bottleneck (that and training the army of pilots that they'd need).

So around 0.4 kWh/passenger-km? - seems it compares pretty favourably to long-haul aviation (e.g. wiuth very crude maths 600 people on A380 at max range/fuel works out as 0.36). Naively I had been assuming that the battery mass and small scale would be more punishing than that.

Even if this will be economically viable, the regulations around it could be so harsh that the idea could not take off.

There have to be rules where those taxis can start, land, in what direction, time frames and over which places in a city. You need some kind of air control and "traffic rules". Drivers need regulated training in simulators and there need to be emergency procedures.

I believe that in the first few years such things won't be allowed near towns and cities. There will be similar rules like flying drones [1], but a lot stricter.

And the time saved by such travel will probably decrease when the sky starts looking like in the 5th Element [2].

[1] https://uavcoach.com/drone-laws-in-switzerland/ [2] https://media.giphy.com/media/Bs2pZhpxf2168/giphy.gif

Initially yes. But the way this starts snowballing is some cities showing they are cooler than others by allowing these things and demonstrating it is fine. Then envy will force the issue and regulation will catch up in most places. IMHO autonomous flight will be the drive this. By the time regulation catches up, having a pilot on board is not going to be a thing any more. Also practically speaking, training that many pilots is just not going to happen in time for that to matter. E.g. deploying 10K of these things would require as many pilots (at least) and they simply don't exist right now. Going from a few hundred to a few thousand will happen relatively quickly but from there to e.g. a few hundred K won't happen until autonomous flight is ready to scale. IMHO we're looking at 15-30 years here for this to happen. With autonomous prototypes flying today, that's a conservative estimate. IMHO the biggest bottleneck will indeed be legal & regulations.

Short term, these things will mean helicopters that currently service rooftops in many cities will be replaced by slightly more of these things flown by professional pilots that will have increasingly less to do as these things start flying themselves. I also expect an increase in heliports. E.g. Manhattan only has a few right now.

I don't see anything fundamentally different than when automobile travel was new. Yes, there will be rules and training, etc. That does not preclude the possibility of having cheap and efficient air travel.

How would this differ from existing regulations for helicopters?

I think you’re going to need to make an actual argument about why the cost would be lower.

Wheels are really awesome at fighting gravity when the engine dies so I’m not sure I ever see a time people are flying around in personal aircraft en masse. I’ve seen enough try to drive on the ground.

Planes and drones are getting pretty good at autonomous flight. You might not need pilots by the time this tech is mature.

They said the same about doctors, and yet heart surgery is done by real people.

All your NNs are entirely completely utterly pointless when faced with inputs they have never been trained on. They always suck at extrapolating, and almost always suck at interpolating as well. So when an unexpected emergency happens, what exactly are they going to do?

And outside of NNs, The only thing we have are autopilots. They sometimes fail, and that is why real pilots mainly just babysit them as their main job.

Just an FYI, planes already regularly fly by themselves. Auto landing has even been around since the late 60's.

Yes but there are still trained human pilots monitoring the auto landing systems, ready to abort if anything goes wrong. It general autonomous systems are incapable of coping with unexpected failure modes.

Sorry, to be clear, planes are not level 5. But to also be clear, there are different levels of autonomy. I'm not trying to suggest they are level 5 though.

The idea that airplanes fly themselves is spectacularly overstated in the public consciousness. The reality is closer to multidimensional cruise control than the pilots are just there for decoration.

Additionally, people forget the extra energy cost of fighting gravity. It takes a lot more energy to use a motor and prop to directly resist gravity, rather than using a static object like a wheel. Energy loss fighting gravity with a propeller >> energy loss fighting wheel friction.

Energy is not the only thing you should care about freight and human logistics. Logistical resources of cities are scarce and it already badly affects labor market there.

Just an FYI, helicopters can land without engine power by autorotation. I'm assuming these things, if you're putting people in them, would do the same.

Additionally, autonomous flight is fairly common right now. There's less going on in the air than on the ground (or rather less surprises with small response time and high consequences). Planes even land and take off by themselves. The first plane to actually be certified to autoland was in 1968[0]. The tech has been around for awhile and been making serious development this entire time.

[0] https://en.wikipedia.org/wiki/Autoland

Those powered lift aircraft don't have enough rotor inertia to perform autorotation. They will have to rely on redundant power systems plus ballistic recovery parachutes.

Autonomous flight is still quite rare. And even when it is used, there are still human pilots standing by to take over in case anything goes wrong.

> Autonomous flight is quite common. When it is used, there are still human pilots standing by to take over in case anything goes wrong.

FTFY.

I'm just saying that comparing autonomous flight to autonomous driving isn't fair. Flying is a lot easier for a machine. You already have tons of data streaming in from airports, other planes, and lots of equipment on the vehicle itself.

That's completely wrong. Sure machines can control take off, navigation, and landing as long as everything works. However they are unable to cope with unexpected system failures. A lot of things can go wrong on even the simplest aircraft and it's impossible to anticipate and code for all of them.

There's different levels of autonomy, just to be clear. I'm not saying they are level 5.

From what I gather most companies in the space plan to land and take off from rooftops. That leaves a lot more options for stations and edges connecting them leading to a denser network than something like a subterranean train which requires dedicated stations with tunnels in between them

In that case, what makes them more useful than helicopters?

They aim to be much cheaper to run. Helicopters are mechanically complex so require a lot of maintenance, and have multiple single points of failure so you can’t skip the maintenance, and are demanding to learn to fly. The hourly rate of a helicopter is high.

Electric motors are simple and reliable and all these companies use multiples so have some redundancy. Multi-rotors require computer control for stability. With that done the piloting is easy.

Sure, they're trying to make it cheaper. But all of the eVTOL designs I've seen can't autorotate, making them much less safe (contrary to your implication about redundancy).

Everything I've seen about this space is people with big checks trying to cut corners to enter a new market, and it seems ripe to fail much like the 737-MAX.

https://www.forbes.com/sites/jeremybogaisky/2019/12/01/insid...

And that's ignoring all of the problems with the market itself, not to mention regulations, as well as logistical headaches of transporting between rooftops (which by the way, many cannot safely support landing pads - so now you also have to retrofit skyscrapers with incredibly expensive infrastructure).

To be fair you answered my question and I'm just being belligerent, so thank you for the response.

The argument is that you don’t need to autorotate if you can land under control on 2/3 of your 16 motors and the failure rate of each is epsilon.

The likelihood of failure is much lower with an eVTOL than a conventional helicopter. Additionally a BRS chute could work with eVTOLs in a way it couldn't with a normal heli.

> The likelihood of failure is much lower with an eVTOL than a conventional helicopter.

You're basing this on a myriad of assumptions as eVTOL designs are barely at the prototype stage right now. Helicopters have been in use for decades, so we know how safe they are.

If they truly prove to be safer then that's awesome, but based on everything I've seen I don't believe that to be true.

Why don't you believe that to be true? There are a myriad of ways for a combustion powered helicopter to fail. Just look at how many billionaires have been killed in helicopter flights.

Electric motors do not fail as often as combustion engines. Most eVTOL designs have redundancies that don't exist in a normal helicopter. I have flown R22s a handful of times and there were many times where an engine failure would have been disastrous.

What makes you think people aren't working on all the problems you think they are ignoring? I personally know at least one person working each and every one of the problems you mentioned.

Just because they're working on the problems doesn't mean they're going to solve them. When you look at all the pieces and read that article about the clusterfuck that is Kitty Hawk, you start to get a sense of just how much buzz and how little substance there may be.

FWIW I've interviewed at many of the eVTOL companies and found too much sizzle and not enough steak.

But to each his own. Ultimately it's just my opinion. Reality will assert itself regardless of what I think.

They are all-electric, 100x quieter, notably safer, and significantly cheaper by the mile.

There's actually examples of this in the world now. In São Paulo, the rich take helicopters between buildings [1].

[1] https://www.theguardian.com/world/2008/jun/20/brazil

Even so there are only 469 (in 2008 according to the article) in all of São Paulo and that’s the metro area with the most helicopters in the world.

I doubt the likes of Hyundai are looking to sell their things “in the hundreds to thousands” range.

It’d a a big change in air traffic and noise pollution.

You'd be surprised how far 469 heliports go. That's one more than the number of stations the NY Metro has, which is the metro with the most number of stations in the world. Now imagine you could travel as the crow flies between any station in NYC. That's awesome efficiency, and if you still need you can grab an Uber at street level if you're still not close enough.

You're also not limited to just those 469 heliports. The vertices of the convex hull they produce can all serve as jumping off points to go about a 100 miles further to heliports one may have in ones own home.

For the rich in São Paulo, that may be homes in places like Angra do Reis and Laranjeiras. Even places like those two (which sit between Rio de Janeiro and São Paulo, can serve as connection points to jump from São Paulo to Rio de Janeiro with only one or two stops in between the two cities.

When you eliminate the time in traffic traffic to get to Congonhas or Guarulhos and the time spent in airport security, it makes a ton of sense for those with the means. Eventually as the tech progresses and becomes cheaper it will become accessible to more than just the very very rich.

No. Seoul Metro has nearly double that of NYC. I have no idea how you could possibly think NYC has the most in the world. Just think of most Chinese cities, let alone Korea's capital. https://en.wikipedia.org/wiki/Seoul_Metropolitan_Subway

Also Tokyo, Japan. It's ranked first in the world on subway usage and has 882 railway stations. https://en.m.wikipedia.org/wiki/Tokyo_subway

Actually Tokyo has much less than Seoul. You're counting railway stations, not subway stations. There's a large difference. Though the water is quite muddled, the easiest way is to actually look at the maps of the differing cities.

I was actually very let down by the metro of Tokyo compared to other Asian cities. Probably due to my high expectations.

Thanks for the correction. Google failed me when I tried to look up the metro with the most number of stops.

that smog tho...

They also used to do this in NYC a lot. Here's a Bloomberg video on it that also discusses why it failed https://www.youtube.com/watch?v=8nbz5VFilxY

Except those rooftops are very expensive real estate, that often already have other stuff on them. So to rent one, you need to pay a bunch of money, which increases price per ride and decreases total rides... that combo means this isn’t that realistic for wide spread adoption.

do these have clear benefits compared to choppers ? open question.

Half of these designs are simply scaled up drones which, according to simple physics and engineering, is the wrong direction.

Details: http://datagenetics.com/blog/february62019/index.html

How can organisations this large get it so wrong? When is someone going to tell them that the Emperor is not wearing any clothes?

I think multi-rotor is a great safety feature, especially if they're partitioned into separate power systems. It would almost require intention to design something like the Volocopter without the ability to soft land in the case of an emergency.

The link you provided mentions this in an odd way:

> With a single engine helicopter you don’t have this redundancy. However, on engine failure, the single massive disk of the helicopter is an asset. ... Close to the ground, this stored kinetic energy can be traded, through pulling up the cyclic, to allow a safe landing. This is called autorotation, and is practiced by all helicopter pilots.

The air taxi that succeeds, at any sort of scale, will not be meant or designed for skilled helicopter pilots pulling levers, at the right moment, as they speed towards the ground. They will be dumb, cheap to service, unskilled to fly, and have more than one point of failure.

> pulling levers, at the right moment, as they speed towards the ground

I think this could be automated?

Sure [1], but safety through redundancy is much easier than automation.

1. https://www.youtube.com/watch?v=iLIo8jO9VUU

Noise may be a factor in certain locales with decent human-centered regulation against it, but:

A) rich people have a ton of power, they will get it legalized if they really want to use them

B) There are many sparsely populated areas where rich people would still be able to make this work...think rural islands of Japan or New Zealand, very mountainous areas, the outskirts of a major metropolitan area...I think these will find use in certain markets.

These are 100x quieter than a helicopter, and nearly inaudible overhead. All-electric range is 60 miles at 150mph, which covers the vast majority of popular and congested urban routes.

The people with the money for these have the political power to force others to deal with the noise. Already, helicopters have remarkably few restrictions. If you can afford one, you can fly it almost anywhere you want.

Helicopters are so expensive to own and operate that the number of people that are even in the running to cause noise is low to begin with.

Exactly, people with that much money can lobby against any noise restrictions.

I really don't see a pattern of any sort of large scale noise pollution via helicopter forced on people to think it will somehow be sure a thing.

It’s not large scale because it’s super expensive. You don’t have a problem with helicopter noise because they are almost no helicopters.

Reply on news.ycombinator.com