Hacker News Reader

Comment by imgabe | Hacker News Reader

I hope they fly themselves as well. Humans can barely manage to safely pilot a car in 2 dimensions. I don't want to think of the carnage that would result if we took existing traffic and moved it 10,000 feet up in the air. Every fender bender would become a multiple fatality.

View on HN

> if we took existing traffic and moved it 10,000 feet up in the air

Wouldn't you move it 1,000 feet up, 2,000 feet up, 3,000 feet up...

There's a lot more space in the air than on land.

Plus, in 3D you can make two non-intersecting non-parallel straight paths. Can't do that in 2D.

> There's a lot more space in the air than on land.

But that's mitigated by the fact that the failure modes and much, much worse.

I can drive safely at highway speeds with a couple hundred feet of separation between cars.

FAA rules on safe separation between planes laterally is about 3 miles, or about 50 times the distance, or 2500 times the surface area.

Plus planes compete with one another for access to airports, so you haven't actually solved any congestion issues -- the bottleneck is still near the ground.

A 3 mile horizontal radius has little to do with flying and a lot to do with speed.

At ordinary commercial airline speeds, it's the equivalent of a "20-second rule" (i.e. it's the distance covered in 20 seconds by a 500-600mph vehicle).

There is a 250KIAS speed limit below 10000'MSL that applies to everyone flying VFR including airlines. They're quickly climbing out above that and end up being a non-factor. The real issue is all the other general aviation aircraft.

Just the same, cars can be driven almost bumper to bumper, at arbitrarily low speeds, in ten foot wide lanes. The realities of wind and minimum forward speed and prop wash dictate that aircraft will need a lot more clearance than a car does. I'm sure it will depend on the particular air vehicle and the rules of traffic in the sky, but I for one would have to be convinced that aircraft can give better throughput than a road can. On the face of things, I'm not sure that's the case.

That's for visual flight rules. There's no way this works with VFR. For one, the examples are all major cities where there's Class B airspace, requiring positive separation. So it's basically "clear of clouds", however that assumes you're cleared into Class B. There's no possible way this works with the existing ATC system.

Flying cars won't use the same airspace as turbine powered airline planes, so a new airspace designation could exist just for cars. And drones. Because drones and flying cars will be competing for the same airspace so all of them will need to use the same system.

Added: Flying cars (and to some degree drones) do compete with the same airspace as existing general aviation flights. And that's a big problem. The legacy aircraft would have to avoid the new airspace, most won't ever get an upgrade to fly autonomously; alternatively, the flying cars will have to be programmed such that they can always yield right of way to legacy aircraft.

Once you're past the threshold of "the fall will kill you" it's better to be higher.

Gravitational potential energy is a little more reliable than aircraft engines. if you lose propulsion, you can convert altitude to airspeed at your leisure, subject to the glide ratio. But only if you have it :).

Whatever else goes wrong, being further away from the ground gives you more time to deal with it.

I think everyone working on this is well-aware of this problem. It's obvious that there can't be widespread adoption without self-piloting technology. The danger to passengers and people on the ground is much too great, and the burden of getting a pilot license is also too much of an obstacle for most people. There are other considerations such as requirements for frequent routine inspections on flying vehicles, which many people couldn't be trusted to fulfill, which also suggest that a self-driving-uber kind of model is really the right solution.

Shameless plug: my blog post from last year on this topic: https://pointersgonewild.com/2015/09/15/all-hope-is-not-lost...

Automatic flight is much easier than autonomous driving. Not much to run into up there :).

Until everyone has a flying car?

>Not much to run into up there :)

So far. In a future where people commute in flying cars while delivery drones descend from some flying warehouse I'm not so sure this stays true.

The saving grace is probably that up there we don't have to accommodate legacy vehicles. We can just force all airborne vehicles to talk to each other with some universal collision avoidance/traffic routing protocol.

> We can just force all airborne vehicles to talk to each other with some universal collision avoidance/traffic routing protocol.

It's not exactly what you described, but in the US, the FAA already has a requirement for aircraft to broadcast their position and velocity in all but the least congested airspace by 2020 [1]. This has been in the works for over a decade [2], and among other things, the information provided by this system can be used to augment existing collision avoidance schemes [3].

[1] https://www.faa.gov/nextgen/programs/adsb/faq/#o2

[2] https://en.wikipedia.org/wiki/Automatic_dependent_surveillan...

[3] https://en.wikipedia.org/wiki/Traffic_collision_avoidance_sy...

Are you familiar with Smeed's Law? I think that it will also apply to flying cars. https://en.wikipedia.org/wiki/Smeed's_law

I doubt it, there's just so much space up there. So first, no concept of lanes. Thinking in 2D space, everything becomes a road. Increase that another dimension, so now everything is a road and there's hundreds of levels of road.

As long as the flight software keeps the cars/planes a decent distance from each other traffic will never be an issue when flying.

A much bigger problem will be parking, since no traffic means everyone will fly everywhere all the time. Parking will be an extreme issue for city centers. The wait in traffic will probably turn into a wait to park

I suspect that, even in 3d space, there will be a concept of lanes. You still want volumes of space designated for one direction or another, at the very least to reduce the number of collision-calculations the guiding system has to make.

Parking shouldn't be an issue? We'll have vertical parking structures that can be entered/exited on every level.

I wouldn't want to fly any kind of helicopter or multicopter into a door in the side of a traditional multi story parking building. the sudden and asymmetric entrance into ground-effect will roll the aircraft, and I'm not sure what having a roof above your rotors would do - possible choke them of airflow and force them into high-speed-stall.

airflow near walls is turbulent, and the funneling effect may suck the aircraft towards the wall.

Solvable problems - you just have to build a landing pad away from the side of the building.

Totally true. I imagine you'd have to have some kind of checkerboard of landing pads.

And, if our flying cars are helicopters or multicopters, we've got more problems, because we can't actually stack that many helicopters over eachother, due to the interference of their air columns. :-/

I think realistically all of the caroplanes will be built more like quadrocopters than normal aircraft. It's much less efficient but the only thing that really meets manuevability and safety constraints

In general aviation (anything other than the airlines for our purposes) midair collisions happen with alarming regularity anywhere there is any congregation of aircraft (typically within 10mi of a non-towered airport). Add to this that they are nearly 100% fatal, unlike traffic collisions, and you have a recipe for disaster. There are systems coming online to help mitigate this issue (ADS-B) but they've been in the pipeline for years and won't see universal adoption until 2020 in theory.

"Not much to run into up there :)"

Birds, hills, turbulence.

Ya. Having just taken a 6 personal commercial plane in stormy, windy weather I'm glad I had two professional pilots driving. It was absolutely terrifying, and I can't imagine your ordinary person feeling comfortable with it. I couldn't wait to get off the plane ASAP and thought we might crash on landing.

Honestly the plane was likely flying itself 90+% of the time. The computers are much better and giving a smooth ride in turbulance.

For fun take a lot at the outer trim tabs when flying in rough skies. The computer adjusts them like 20 times a second to keep the wings stable, kinda cool to see

Other flying cars.

Surely it would be simpler than dealing with the complexity current '2d' autonomous car systems face?

As long as you have designated drop zones, everything else becomes mathematically simpler.

But their entire path would require a drop zone. So in effect, there would be these long strips on the ground mirroring the flight paths of these flying cars. You might even run out of space on the ground for drop zones.

With the type of aircraft in the article, they could come down like helicopters do, right?

I figured that the drop zone would be used if the vehicle encountered some failure, and had to land immediately (or crash land even), which would prevent it from continuing flight to a smaller drop zone that's some distance away.

Just follow freeways. Though that gives up point to point efficiency.

Just designate all (currently free) parking spaces as drop zones, and all roads as emergency drop zones.

Since ground-based cars will in all probability continue to be a thing, roads are not guaranteed to be empty, but are guaranteed to be plentiful and nearly everywhere. Autonomous cars can be signaled as some road strip is reserved for emergency landing, and within seconds they can vacate the road strip and block it from conventional vehicles (assuming a sufficient portion of ground vehicles becomes autonomous).

Roads are guaranteed not to be empty enough for this purpose in any urban area - road traffic might diminish somewhat, but not go away, with the arrival of flying cars.

Even if a flying car could fold up to be the size of a current car while parked (which is not at all an easy matter) it will not be that small in a landing configuration.

It's ok, autonomous flying cars will save the day

Reply on news.ycombinator.com