I've seen a video of a computer controlled plane landing, taken from inside the cockpit. The pilot got some serious heebie jeebies, but the landing was fine.
Yeah that's the old Airbus crash thing, very old, pretty dramatic. The pilot was flying too low and the aircraft computer thought it was coming into land, with the inevitable results..
It's so horrible, so much death... at least I assume a lot of death...
And yet the stately way it flies into the trees, nose in the air, is hilarious. It's like watching a British butler solemnly and deliberately walk into a wall.
This is more of a comment about other comments here than a specific response to Giza's post about the problems Airbus had early on with fly by wire.
To an extent, yes, computers frequently fly airplanes. It really depends on what one considers 'flying the airplane'. In an Airbus (except one), there is no direct connection between the pilots and the aircraft. Computers take pilot input as a strong suggestion, then decide how to actually manipulate aerodynamic control surfaces to maneuver the aircraft. Most have heard of this as "Fly By Wire."
Outside Airbus, and with notable exception of the Boeing 777, "computers flying the airplane" could be written "computers simulating pilot input to the controls". An autopilot in most commercial aircraft takes information from navigation computers and drives servos to mechanically control the yolk (stick) as if a pilot was physically manipulating the yoke. The same goes for autoland computers. At least in these systems, if the computer(s) start going on the fritz, the pilot has the option to regain physical control of the aircraft. Fly by wire does not give this option.
For the record, computers (autopilot or otherwise) do very little of my flying except in long range cruise. I prefer to do everything else myself.
(no subject)
(no subject)
Date: 2006-03-03 07:52 pm (UTC)(no subject)
Date: 2006-03-03 08:28 pm (UTC)(no subject)
Date: 2006-03-03 08:39 pm (UTC)What Elfasi said, pilot error on landing (My mistake, was thinking another simular event).
Lots of planes fly the world 99% of the time controlled by computer and there's never been a major event concerning them.
(no subject)
Date: 2006-03-04 12:50 am (UTC)(no subject)
Date: 2006-03-04 05:38 am (UTC)And yet the stately way it flies into the trees, nose in the air, is hilarious. It's like watching a British butler solemnly and deliberately walk into a wall.
(no subject)
Date: 2006-03-04 05:34 pm (UTC)*CRASH*
Not anymore. =)
Computers flying airplanes
Date: 2006-03-05 02:47 am (UTC)To an extent, yes, computers frequently fly airplanes. It really depends on what one considers 'flying the airplane'. In an Airbus (except one), there is no direct connection between the pilots and the aircraft. Computers take pilot input as a strong suggestion, then decide how to actually manipulate aerodynamic control surfaces to maneuver the aircraft. Most have heard of this as "Fly By Wire."
Outside Airbus, and with notable exception of the Boeing 777, "computers flying the airplane" could be written "computers simulating pilot input to the controls". An autopilot in most commercial aircraft takes information from navigation computers and drives servos to mechanically control the yolk (stick) as if a pilot was physically manipulating the yoke. The same goes for autoland computers. At least in these systems, if the computer(s) start going on the fritz, the pilot has the option to regain physical control of the aircraft. Fly by wire does not give this option.
For the record, computers (autopilot or otherwise) do very little of my flying except in long range cruise. I prefer to do everything else myself.