Does it occur to anyone that the folks who are working on driverless cars might have a reason to want a place to work on pilotless planes? While the current autopilot in many commercial planes can handle a lot of the flying, there are some decent savings available if you could reduce the number of pilots. Especially if you can make the solutions small and cheap enough for 2-4 seat planes.
back
6 comments
My understanding of modern airlines is that unless your plane is in the process of taking off, landing, or crashing, it's already essentially pilotless.
There's always been too much at stake in the airline industry to allow a lot of room for human error. Combined with the huge leaps that fly-by-wire planes have made in the last few decades, that made for some very powerful incentives on everyone's part (engineers, regulators, and companies) to advance autopilot as far as they possibly could.
In normal conditions, yes. But when things go wrong (your "or crashing", perhaps), the pilot is suddenly very much involved, and screwups can be fatal.
Take, for example, Qantas flight 32, an A380 that was successfully landed after one of the engines more or less blew up, damaging several other systems. Truly pilotless operation would require automated systems that could do as well --- and hopefully better than the pilots of Air France 447, the transatlantic flight that was lost due to recoverable pilot error.
Yes. The failure/recovery mode for "speedometer is reporting inconsistent speeds" is stop the engine and coast to a stop. That doesn't work as well with planes.
Then again, human pilots are fallible, and lots of crashes would be avoided if they paid more attention to instruments and less to instinct (where's that horizon again?), so maybe programming planes to just blindly follow the instruments would result in a net reduction in crashes.
While your last sentence may be true-ish (I'm not really qualified to answer with authority), your first sentence is very much not.
This relevant link gets posted a lot, but is worth posting again: http://www.salon.com/2011/08/04/can_jetliners_fly_themselves...
Your link doesn't argue with Cowen's first sentence. When Cowen says "essentially pilotless" I understand it to mean "keeps going without operator intervention", and after reading your link that's still my understanding of what happens.
I'm not sure you and I read it in the same way:
And what do terms like “automatic” and “autopilot” mean, anyway? Autopilot is simply a tool, along with many other tools available to the crew. You still need to tell it what to do, how to do it and when to do it. And although it frees the pilot from having to have his or her hands physically on the control wheel (or sidestick) or thrust levers, this is only a fraction of what the act of “flying” entails. The autopilot is not flying the plane. The crew is flying the plane, through the autoflight system.
That in no way says "keeps going without operator intervention" to me.
The technology is pretty much there, but currently there's still a requirement for human supervision/control in case of mechanical or electronic failures.
Like a need to land the plane in Hudson River when the plane hits a flock of geese, or emergencies when vital system monitoring tools become inaccessible.
A large portion of FAA's required reading for pilots http://www.faa.gov/library/manuals/aviation/pilot_handbook/ is dealing with various failures of subsystems.
Good luck pitching that to the FAA. I think they'd happily shut down aviation before they'd let planes without pilots carry passengers. A human being with his life at stake is extraordinarily resourceful. Sure, 'pilot error' is the cause of quite a few incidents. But 'pilot lands plane in difficult circumstances' outnumbers pilot error by a substantial amount, it's just that those don't get nearly as much press.
So I don't think that Google is working on 'pilotless planes'.
To continue the satire from the parent post...
Yes, remotely piloted or autonomous flying drones could never work. There must be 2 pilots flying a bird, that will never change, even the Air Force wouldn't be that crazy.
But honestly, give it twenty years.
Airplanes are already capable of operating without pilots much better than cars. The only remaining issue is listening to flight controllers, but that's a bloody difficult problem.
Isn't that the easy part though? The onboard pilots will automatically listen to flight controllers.
But honestly, don't you see the solution here? You have a room full of pilots in Tulsa Oklahoma (or wherever) or are providing manual intervention when necessary. Just like how we manage millitary drones today.
Drones are managed very poorly in busy airspace, and tend to crash if there is a link failure. Currently, I think that airline Captains have to specifically set the auto-pilot to fly the final approach and landing, and be ready for a manual go-around if required. Typically pilots use "preset" auto-pilot only for enroute flying to initial hold point, then sometimes use fixed-rate decent with heading hold mode after that, until they turn onto final approach.
The expense and arguments over a next generation of airplane tracking and control are probably going to go on for a while longer.
The savings aren't that big of a deal when you're talking about a flight with 200+ paying passengers. You are still going to have someone in there in case the computer system malfunctions, they might as well be a pilot. Autopilot is driverless flight already.
Savings assuming you’re not a pilot.