how long

[Rockie]

Every flight. The FMS has to be programmed and modified to suit anticipated conditions. Once inflight actual conditions, traffic, poorly performing software requires further modifications or direct control over the flight path of the aircraft either through use of direct modes of the autopilot and auto throttle, or by disengaging them completely.

It has happened on every flight I’ve ever done for reasons you know nothing about.

[digits_]

Every flight. The FMS has to be programmed and modified to suit anticipated conditions. Once inflight actual conditions, traffic, poorly performing software requires further modifications or direct control over the flight path of the aircraft either through use of direct modes of the autopilot and auto throttle, or by disengaging them completely.

It has happened on every flight I’ve ever done for reasons you know nothing about.

Changing autopilot modes is exactly what ATC could do. That's a software command.

Having to disengage them completely is a more serious issue and that would be harder to deal with. What year, roughly, was the airplane built you were flying?

[altiplano]

The fact is nothing short of a fully working general AI will work for removing both pilots completely from the cockpit. And once we get there, EVERYONE Is replaceable (doctors, lawyers, basically everything) so it won’t really matter at that point. They still have drivers in front trains and they’ve been able to automate that for decades.

Why is that a fact?

Most people here seem to assume that pilots can only be replaced if airplanes can operate completely from A to B, without any human intervention at all, and deal with all diversion decision, notams, weather etc.

That's not the case. You still have ATC. What's the difference between ATC telling pilots to change a heading or give them a direct, and have the pilots put it in the autopilot, versus ATC inputting the data directly in the auto pilot? ATC can be the brains if necessary. They'll be on the ground for the tough decisions, and no pilots in the airplane. Thats not sciene fiction, that's just waiting for regulations to allow it, and some manageable technological changes to be executed.

Stick to your web design... you still don't have a clue.

[Zaibatsu]

The most important thing about automation is that it has to be monitored. It’s so important that we have two pilots monitoring it even though none are physically flying it. ATC would be over their head with two aircraft, never mind a dozen.

ATC clears you to climb to 350 and on course. All he knows is that’s what you want and there’s no traffic in the way. He doesn’t know what vertical mode you are going to use to get there. He doesn’t know about the storm cell in your path. He doesn’t know about the CAT you’re going to encounter at 330. He doesn’t know that you’re going to have an emergency or abnormality that’s going to require deviations from your cleared route and quite possibly a change in destination based on first hand information.

And that’s a sleepy area control centre. Not terminal where they are in the middle of switching runways and now you have to reprogram the FMS and rebrief, or the ILS just went offline, or a plane didn’t hold short or occupied the runway for too long. All of those could be dangerous or deadly without direct and present human interventions.

[Rockie]

Every flight. The FMS has to be programmed and modified to suit anticipated conditions. Once inflight actual conditions, traffic, poorly performing software requires further modifications or direct control over the flight path of the aircraft either through use of direct modes of the autopilot and auto throttle, or by disengaging them completely.

It has happened on every flight I’ve ever done for reasons you know nothing about.

Changing autopilot modes is exactly what ATC could do. That's a software command.

Having to disengage them completely is a more serious issue and that would be harder to deal with. What year, roughly, was the airplane built you were flying?

2018.

What part about "every clearance and instruction given to a pilot by ATC can be rejected by the pilot for safety reasons ATC is neither aware of or responsible for" did you not understand?

[digits_]

The most important thing about automation is that it has to be monitored. It’s so important that we have two pilots monitoring it even though none are physically flying it. ATC would be over their head with two aircraft, never mind a dozen.

ATC clears you to climb to 350 and on course. All he knows is that’s what you want and there’s no traffic in the way. He doesn’t know what vertical mode you are going to use to get there. He doesn’t know about the storm cell in your path. He doesn’t know about the CAT you’re going to encounter at 330. He doesn’t know that you’re going to have an emergency or abnormality that’s going to require deviations from your cleared route and quite possibly a change in destination based on first hand information.

And that’s a sleepy area control centre. Not terminal where they are in the middle of switching runways and now you have to reprogram the FMS and rebrief, or the ILS just went offline, or a plane didn’t hold short or occupied the runway for too long. All of those could be dangerous or deadly without direct and present human interventions.

I wouldn't see ATC as a remote pilot for every airplane. They would just issue instructions on their computer screen during standard ops. If for some reason, the "computer brain" or whatever you want to call it, on a remote plane would fail, then they could focus on that one plane.

If you encounter turbulence at a certain flight level, don't you usually query ATC where the better flight level might be? It's easy to add sensors to airplanes to measure the turbulence, the computer brain could then request a climb or descend, or ATC could propose one. That's not hard to build in.

Once you commit to a pilot less aiprlane, the inputs become much easier. No interface is required, no need to confirm critical data for liability reasons, as the computer does it all anyway. A runway chance should be non event at that point. That's easy software to program. The whole IFR system is already set up pretty robust.

During an emergency, you evaluate the systems you have left, the possible destinations and analyze weather, services etc. That's stuff that could be programmed as well. It shouldn't be a problem. Will the computer always find a solution that would be as good as a human would? No. That's part of the trade off humans vs computer that manufacturers, companies and politicians will have to make.

2018.

Ok, if that still happens than your plane would obviously not be a good candidate to fly pilotless at this point. That's why you are still on board. That doesn't mean that a software upgrade couldn't fix these issues.

What part about "every clearance and instruction given to a pilot by ATC can be rejected by the pilot for safety reasons ATC is neither aware of or responsible for" did you not understand?

Nothing, but it is't entirely relevant. ATC now gives instructions and clearances with the understanding that the pilots have the final decision. If there are no pilots on board, it stands to reason that the instructions and clearances that ATC is allowed to give you, will be different.

Another way to think about is this: all the information that is on a screen or a dial in the airplane somewhere, is information that the computer brain has access to. All information that a pilot gets via ATC, is information that the computer brain can also get. What other source of information do you need? In extreme cases where you lose most of your instruments and GPS etc, you might have to look outside to land visually on a runway that is theoretically too short but is your only way out. How often does this happen in an airline scenario?

You can have pax concerns. That can be handled by flight attendants. Press an emergency button, indicating the severity of the pax emergency, and the computer brain will decide where to go.

The technology to do this exists.

[tsgarp]

Airbus demonstrated it, but pilotless drones have been around a while and this was just a bigger version of one.

As you said, the technology has been around, and proven, for a long time. Airbus just took it to the next level. From a technological standpoint, there is nothing preventing pilotless airliners. The determining factor will be the risk assessment combined with the financial assessment. When the probability of a computer caused accident is equal or less than the probability of a human caused accident and when the cost of the computers (and their associated risks) is less than that of a human flight crew it will happen.

#4 is correct. Not quite a million, but it's telling that you take things so literally you can't recognize hyperbole.

I recognize hyperbole quite well, and I accord it, and it's users, the appropriate level of respect as intellectuals.

It is no exaggeration however that every actual pilot that actually flies automated aircraft intervenes constantly in that automation. As I've stated before many times, the FLCH, V/S, HDG and finally the A/T and autopilot disconnect switches are there for a reason and used constantly. According to you the aircraft takes care of itself which shows the spectacular degree of ignorance you live in.

I've flown automation for about the last 15 years or so. Mostly Universal FMSs. There have been two or three times it did something unexpected. Most of those times it did what it did because of something I put, or failed to put, in the flight plan. If you are 'constantly' intervening in your automation because it is doing unexpected things then I would suggest the problem is in the person programming the automation; garbage in garbage out.

You're an engineer right?

No rusty nail on my pinky.

[Rockie]

Tsgarp. Flying automated airplanes for 15 years? Bullshit, you don’t fly anything.

Digits, what do you do for a living?

[Meatservo]

Sometimes I amuse myself (during those long automated cruise segments) by imagining an alternate universe in which technology rolled out differently than it did in this one. I like to imagine a world where aeroplanes have always been automated, ever since they were invented. Like for some reason, electronics were invented before flying machines. Every aircraft since the dawn of flight has been crewed, but has been controlled exclusively via an FMS and mode-control panel. Analog controls would be provided in the form of a bank of little knobs that individually selected heading, angle of bank, rate of climb, altitude and speed. THEN one day, some genius invents something. Maybe it's originally for astronauts, to make docking with a space-station a bit less effort-intensive, or whatever. But it trickles down into all levels of aviation. It consists of a stick-like handle, a pair of pedals, and a lever, or levers I guess, that directly modulates engine output. The stick-like device combines pitch and roll commands into one control. The pedals allow the pilot to yaw the aircraft. It would be called the "universal haptic control interface". It would be revolutionary. Suddenly a human operator would be capable of commanding multiple flight-attitude adjustments at once, simply with small movements of his limbs. Airbus would be the first to adopt the technology. Pilot unions would resist it because of the re-training involved. CEOs would love it because they could save money and weight on all the electronic stuff. Advocates of the U.H.C.I. would compile all kinds of testimony and damning reports concerning the times A.I. fucked up and a direct interface would have allowed a crewmember to intervene more quickly. Pilots themselves would be divided on the subject. Overall it would be quite controversial.

[digits_]

Sometimes I amuse myself (during those long automated cruise segments) by imagining an alternate universe in which technology rolled out differently than it did in this one. I like to imagine a world where aeroplanes have always been automated, ever since they were invented. Like for some reason, electronics were invented before flying machines. Every aircraft since the dawn of flight has been crewed, but has been controlled exclusively via an FMS and mode-control panel. Analog controls would be provided in the form of a bank of little knobs that individually selected heading, angle of bank, rate of climb, altitude and speed. THEN one day, some genius invents something. Maybe it's originally for astronauts, to make docking with a space-station a bit less effort-intensive, or whatever. But it trickles down into all levels of aviation. It consists of a stick-like handle, a pair of pedals, and a lever, or levers I guess, that directly modulates engine output. The stick-like device combines pitch and roll commands into one control. The pedals allow the pilot to yaw the aircraft. It would be called the "universal haptic control interface". It would be revolutionary. Suddenly a human operator would be capable of commanding multiple flight-attitude adjustments at once, simply with small movements of his limbs. Airbus would be the first to adopt the technology. Pilot unions would resist it because of the re-training involved. CEOs would love it because they could save money and weight on all the electronic stuff. Advocates of the U.H.C.I. would compile all kinds of testimony and damning reports concerning the times A.I. fucked up and a direct interface would have allowed a crewmember to intervene more quickly. Pilots themselves would be divided on the subject. Overall it would be quite controversial.

STAR TREK!
https://en.wikipedia.org/wiki/Extreme_R ... :_Voyager)

Ok, only a small part of that episode was about the manual controls, but still, never let an opportunity to reference Star Trek go to waste! Especially if it concerns Voyager...

['97 Tercel]

My prediction is that there will continue to be human pilots in the front of airliners as well as increasingly capable auto pilots, until such time as they invent a system that is sufficiently advanced to function as an auto-SCAPEGOAT.

Exactly.

Captains are just Risk Managers and until they get a suitable replacement(30-40 years?) at least one human will be in the cockpit

[SpyPilot]

Never let an opportunity to reference Star Trek go to waste, especially TOS

https://en.wikipedia.org/wiki/The_Ultimate_Computer

[Rockie]

My prediction is that there will continue to be human pilots in the front of airliners as well as increasingly capable auto pilots, until such time as they invent a system that is sufficiently advanced to function as an auto-SCAPEGOAT.

Exactly.

Captains are just Risk Managers and until they get a suitable replacement(30-40 years?) at least one human will be in the cockpit

The Aeronautics Act disagrees with you.


"pilot-in-command"
«commandant de bord»
"pilot-in-command" means, in relation to an aircraft, the pilot having responsibility and authority for the operation and safety of the aircraft during flight time;


Replacing "pilot" with "circuit board" isn't going to happen as long as human beings are involved because no one will want to give it authority over their life, and you can't hold a circuit board responsible. Why do you think cars with autopilots have all those warnings? The driver is ALWAYS responsible.

I've asked this before as well and still no one has ventured to answer it, if it goes to a single pilot who is going to train them?

[valleyboy]

Most are thinking "inside the box" and looking mostly at the present and short term future. If you don't believe that
it will evolve into completely automated flight that that the only human input will be destination and departure time you are restricting your mind. WX will have little effect except in situations such as hurricanes and thunder storms for arrivals and departures but pilots not needed to deal with that. Enroute will be above all WX and low orbit and as I said before short travel will be ground based. Come on guys lets face it computers can fly much better and more efficiently that any human can or could. The tech by then will be solid and all traffic (except for the general aviation hobbyist) will be controlled by a world central computer. Ironically there will likely be fewer aircraft in the air at any one time. Humans directly operating machines will become non-existent. The only thing that might change this is we blow ourselves up before it evolves and yup by then all humans will be wearing body condoms with re-breathers

[trey kule]

Rockie asked..

“I've asked this before as well and still no one has ventured to answer it, if it goes to a single pilot who is going to train them? “


That is definitely a challenge. Particularly, as no where in the world right now is any plane being flown single pilot. I , for one, admit that I never have seen a single cockpit aircraft, or a plane being flown by one pilot. A training issue. Who knew?

Maybe some smart people could develop, oh I don’t know....some kind of simulation device.
Or maybe make a two person trainer. I know. I know . Far out ideas, like making a simulator for something like a space shuttle. Can’t be done. Simply impossible to train someone for something that has not been done before.

A question for the ages...Who trained the first pilot?

[Rockie]

That is definitely a challenge. Particularly, as no where in the world right now is any plane being flown single pilot. I , for one, admit that I never have seen a single cockpit aircraft, or a plane being flown by one pilot. A training issue. Who knew?

Maybe some smart people could develop, oh I don’t know....some kind of simulation device.
Or maybe make a two person trainer. I know. I know . Far out ideas, like making a simulator for something like a space shuttle. Can’t be done. Simply impossible to train someone for something that has not been done before.

A question for the ages...Who trained the first pilot?

Well, having trained many very experienced pilots in these airplanes including many new Captains I can confidently say you are full of shit, making the completely uninformed assumption these aircraft or the job are anything like boring holes in the circuit with your Cessna 172.

[Rockie]

WX will have little effect except in situations such as hurricanes and thunder storms for arrivals and departures but pilots not needed to deal with that.

I once knew a pilot who wanted an airplane that could do 400 knots because "it will outrun all the weather".

[Meatservo]

That is definitely a challenge. ..Who trained the first pilot?

Well, having trained many very experienced pilots in these airplanes ......, making the completely uninformed assumption these aircraft or the job are anything like boring holes in the circuit with your Cessna 172.

redacted to eliminate rudeness

Well I suppose the point is that automation could theoretically reduce the workload until an airliner really is no more challenging to operate than a Cessna 172. I am reminded of the humble Cessna Caravan, which operates single-pilot IFR in much of the same low-level controlled airspace as arriving and departing airliners, and has enough rudimentary automation and system-simplicity to allow it to be operated without too much trouble by one pilot. Maybe the somewhat less humble Pilatus PC-12 is a better example. Heck, even the new Pilatus PC-24- there you go- Which supports "Trey Kule"s point of view.

Alternatively, the aforementioned Cessna Caravan has a horrible safety record if you applied it to the 705 heavy-people-mover world. Partly because of pilot inexperience. In a world where there is ONLY one pilot, what would be the minimum qualification necessary? Rather than simply ask "who will train them", it might be more correct to wonder what kind of career progression will occur in this professional scenario. How many hours will a new-hire have? What kind of experience will they need to have? Where will they get this experience? Being able to hire experienced flight-crew without relying the probationary period of having them act in a subordinate position would seem to require that you have a pool of pilots who have gained that experience somewhere else. Would there be enough "somewhere elses" to supply the demand? These questions support "Rockie"'s point of view.

Even now the question of experience bears examination. Many pilots who are convinced that human pilots are still relevant like to talk about the automation failing, and the experience and raw talent of the flight-crew being the lifesaving final solution. So when we get to the point (arguably in some cases we're there already) when the sum total of a flight-crew's experience consists of operating within the framework of a highly automated environment, from which orifice will they produce this "skill" and "experience" that will save their lives if/when the automation fails?

I think all these questions make for interesting debate.

[valleyboy]

I once knew a pilot who wanted an airplane that could do 400 knots because "it will outrun all the weather".

I would rather have low earth orbit aircraft which is well above the wx and certainly makes 400 kts look stationary

[Zaibatsu]

I wouldn't see ATC as a remote pilot for every airplane. They would just issue instructions on their computer screen during standard ops. If for some reason, the "computer brain" or whatever you want to call it, on a remote plane would fail, then they could focus on that one plane.

If you encounter turbulence at a certain flight level, don't you usually query ATC where the better flight level might be? It's easy to add sensors to airplanes to measure the turbulence, the computer brain could then request a climb or descend, or ATC could propose one. That's not hard to build in.

Once you commit to a pilot less aiprlane, the inputs become much easier. No interface is required, no need to confirm critical data for liability reasons, as the computer does it all anyway. A runway chance should be non event at that point. That's easy software to program. The whole IFR system is already set up pretty robust.

Remember we have an airplane that’s been grounded for over a year because one tiny aspect of automation wants fly it into the ground against the wishes of the two pilots.

There’s nothing very easy about what you suggest.