5 ms·

> Having every random person fly around is an absolute regulatory and organizational nightmare I wouldn't be surprised at all if one day autonomous flight for

> Having every random person fly around is an absolute regulatory and organizational nightmare I wouldn't be surprised at all if one day autonomous flight for low cost aircraft like this one would turn out to be much easier than autonomous drive for cars. I mean the computational problems behind it; energy requirements are of course a whole different story though, and will keep prices very high for quite some time, but I can envision a future in which autonomous flying aircraft can talk each other to avoid collisions and/or arrange taking off or landing from/to shared reduced space.

Mo3 4y ago
I agree, but that is way into the future. Cars are not nearly able to do this, and they only operate in two dimensions. Aircraft operate in three dimensions and are affected by more factors than cars such as wind and weather.
- more_corn 4y ago
  Autopilot for air is probably far easier. Take off, land, fly straight. Begin listing the functions for autonomous car. I’ll wait.
  - Mo3 4y ago
    I’m not sure about that. Both seem quite complicated. Aircraft have different problems and are definitely not as easy to automate as you claim here
  - 3np 4y ago
    Start, drive straight, turn, stop.
- squarefoot 4y ago
  > Cars are not nearly able to do this, and they only operate in two dimensions. True but they depend on roads that are built all different from each other, full of irregular shapes and surprises, bumps, pedestrians, bicycles, absent, broken or different signs when traveling to other counties, etc.
nradov 4y ago
Autonomous IFR flight is a reasonably tractable technical problem, as long as nothing breaks. The Garmin Autonomi system can already sort of do that for light aircraft under some very limited circumstances. The problem is that there are a zillion different potential failure modes (some involving circumstances completely outside the designer's control) and it's impossible to program in responses to all of them. Where experienced human commercial pilots really earn their pay is in handling those rare emergencies.
throwaway14356 4y ago
launching people with cannons seems both cheap and fast

Mo3 4y ago

I agree, but that is way into the future. Cars are not nearly able to do this, and they only operate in two dimensions. Aircraft operate in three dimensions and are affected by more factors than cars such as wind and weather.

more_corn 4y ago
Autopilot for air is probably far easier. Take off, land, fly straight. Begin listing the functions for autonomous car. I’ll wait.
- Mo3 4y ago
  I’m not sure about that. Both seem quite complicated. Aircraft have different problems and are definitely not as easy to automate as you claim here
- 3np 4y ago
  Start, drive straight, turn, stop.
squarefoot 4y ago
> Cars are not nearly able to do this, and they only operate in two dimensions. True but they depend on roads that are built all different from each other, full of irregular shapes and surprises, bumps, pedestrians, bicycles, absent, broken or different signs when traveling to other counties, etc.

nradov 4y ago

Autonomous IFR flight is a reasonably tractable technical problem, as long as nothing breaks. The Garmin Autonomi system can already sort of do that for light aircraft under some very limited circumstances. The problem is that there are a zillion different potential failure modes (some involving circumstances completely outside the designer's control) and it's impossible to program in responses to all of them. Where experienced human commercial pilots really earn their pay is in handling those rare emergencies.

throwaway14356 4y ago

launching people with cannons seems both cheap and fast