I'm starting to think automation may be the answer

[Zionstrat2]

As a result of the Vilnius and Toronto accidents, I'm starting to think that total automation of commercial aviation is inevitable.

This comes as a complete surprise, as I'm a rather limited GA pilot with no automation experience. Everything I love about flying comes from hands on the yoke, and for many years, the magenta line discussion seemed to make it obvious that more hands on experience has been necessary.

However, these two events would have been unlikely with total automation and for the first time I think I get the single pilot idea.

Let George fly to reduce costs and increase safety with a backup pilot simply there to take control in a complete emergency.

A backup pilot wouldn't need to be able to deal with subtle issues like unexpected gusts or even hitting the wrong switch. Once an emergency was declared, I assume an entire network of people on the ground would come together to put the pilot in the best possible circumstances for a safe landing.

It definitely doesn't sound like fun, but I wonder if it's inevitable?

There may be a close analogy with family farming in the United States. Anyone who wanted to farm could until it became unprofitable due to the benefits of scale, limited market access and, of course, large scale automation.

Kind of like me in a 172, a family farmer can still farm. But they can't compete in the big leagues on any level from ROI to safety.

So maybe I'm really late to the game... but is total automation inevitable for commercial aviation? And if so, how soon is it getting here? Obviously I'm not a guru here so very interested in your thoughts.

I don't post here very often, so please forgive if I've posted in the wrong forum and feel free to move it.

[GlobalNav]

What could go wrong, go wrong go wrong….

[rudestuff]

A nice thought, but we're not ready yet. We need the warning, navigation and communication systems in place first. TCAS is inhibited at low altitudes, so we're relying on looking out (human) and ATC radar (human/communication) Radars can fail. ATC can screw up. Communication can be lost. GPS can be jammed or worse - spoofed. Many systems can be inop and the flight can legally dispatch according to the MEL - leading to reduced redundancy, all of which needs to be managed by someone or something to within acceptable levels of risk. Automation is great when it works, but every single day airliners lose comms, lose radars, lose TCAS, Rad alts, FMCs, Autopilots, GPS etc, and are safely landed by humans without you ever hearing about it. Due to the glacial rate of evolution within aviation - particularly in regards to certification - pilots are going to have here for quite some time.

[Cough]

Vilnius - Automatics were engaged and were still available (autopilot A I believe would have worked) after B hydraulics were turned off. Would the aircraft have crashed anyway if the automatics were engaged? If you clear the warning system for the hydraulic failure and don't monitor flap deployment then even with the AutoPilot engaged, the outcome is uncertain.

Toronto - Was the aircraft fitted with AutoLand? Would it have been within wind limitations to autoland even if it was fitted?

I get where you are coming from, the automatic system designers have to be confident that they've thought of EVERY possible outcome before the automatics can take over. Thing is, this just ain't possible as aviation whilst mostly certain, is subject to threats which can put you outside of the design envelope for the systems. We don't have statistics for the number of times pilots have saved an aircraft load of people due automation failure (it's called just doing the job - report it to maintenance afterwards). If we did, I suspect this thread would cease to exist.

[Junkflyer]

Automation is great.
Until it doesn't work.

[Zionstrat2]

Thanks, great feedback.

Considering Tesla's driving through painted walls, I'm quite happy that people are still in charge.

You guys brought up one thing I didn't think of whatsoever.. For total automation to work, the ground system would have to be a lot more robust than it is today.

So probably not time to be investing in a total automation startup quite yet :-)

[Someone Somewhere]

Some kind of 'safe return to (air)port' can probably be implemented without too many changes on the ground - it's basically the same as radio failure: start squawking no radio/emergency, send a few CPDLC messages to the same effect, and then line up and land at the most convenient runway and approach expecting ATC to keep stuff out of your way.

Navigation following GPS failure is going to be a reasonably big open question, but I don't think it's insurmountable, especially with a decent number of NDBs and ILSs available.(edit: Better detection of GPS spoofing and sanity/integrity checking of GPS data is essential, regardless of whether automation increases or not. Differential GPS amongst multiple receivers is I think near impossible to spoof without essentially fake satellites, and you accurately know aircraft orientation as a bonus)

Eventually, I would expect to see a system akin to cab signalling/moving block signalling as used on railways. Instead of just giving you vectors or instructions, ATC gives you blocks of airspace and times you can occupy them, with protections that mean it's impossible to clear two aircraft to be in (nearly) the same airspace at (nearly) the same time. Essentially how NATS is handled but you can slice-and-dice it much more tightly if it's automated. Once you have that, it will be much easier to automate normal flight.

The big long-term question is IMHO whether full-authority one-pilot operation is even acceptable due to e.g. Germanwings: it doesn't matter how many redundant computers, IRSs, and radios you have if the pilot can start pulling breakers to force direct law. I think you're going to find that the avionics bays (plural) are sealed in-flight, and the computer will be effectively 'captain' (whether the pilot is flying or not) unless it decides the flight is unsalvageable or it has lost too much capability.

Long term, I think you might see remote control from a ground situation room as the mechanism to deal with severe emergencies or an in-flight upset, instead of pilots on board. There's enough bandwidth to get multiple good video streams and a lot of instrumentation, and you can have ~5 people involved instead of 1-2.

[Zionstrat2]

That's an interesting list... I definitely like the idea of blocking especially if it would allow more direct routes with very little chance of other traffic.

The one big surprise for me would be remote control. I would have guessed there would be too much concern about remote hijacking.

But there's probably a way to encrypt data to the point that it's less of a concern than I imagine.

[Someone Somewhere]

If I was implementing it, from an automation point of view, the answer would be that the aircraft chooses to surrender control. It only does that if the onboard computers feel something catastrophic has failed beyond what can be handled internally. Ground crews would not be able to take control unless the aircraft had already determined that something was very. very wrong (probably well beyond single engine failure).

Significantly larger emergency power would probably needed, even with increased efficiency. I don't think that's actually going to be a major barrier; modern aircraft are sometimes doing fully electric RATs (A350/A380) and I expect large batteries in place of the APU might show up on the next generation airliners.

Ideally I would be saying that an aircraft can remain in 'normal' law with full protections, automation, and continue to autoland following:

• Loss of anything that would already allow an aircraft to stay in normal law and land more-or-less normally, e.g. single engine failure, hydraulic failure, generator failure etc.
• Total loss of a single avionics bay (two firewalled bays is going to be necessary) or cockpit in the event of e.g. fire (that Egypt incident). Ideally, I'd want to be able to land as long as one of bay A, bay B, or cockpit (IFDS and direct law signalling) is intact.
• Total loss of a single class of sensors e.g. all AoA vanes or pitots frozen (XL airways A320, AF447)
  • Loss of all GPS should be feasible.
  • Loss of all IRS will be very hard to handle and would likely be managed by increasing redundancy and multiple suppliers/designs.
• Major fuel leak; detect, locate, isolate, and divert.
• Crew/cabin crew/ATC requests diversion

Reasons to go to off-plane emergency control would likely include:

• No thrust landing e.g. Sully, Gimli.
• Total loss of flight controls e.g. Sioux City, though ideally this will be fixed at the design stage.
• Major instrumentation failure e.g. both pitot and AoA gone or unreliable
• Loss of control inflight (stall, spin, etc.) - an automated airliner probably should never stall so if it does, something is clearly wrong.
• Something else unexpected and unexplainable where it cannot proceed. The Smartlynx A320 training flight is a good example of an 'unknown unknown' though that specific event should not reoccur.
• Edit: gear up and/or flaps up landing
• Edit: fire warnings, engine vibration etc.

FIve people in a room with no space constraints or smoke issues who do nothing but train for emergencies should do a much better job than two pilots who might be in masks.

[Zionstrat2]

Thanks someone somewhere... That's the kind of out of the box thinking I was looking for...

Hadn't entered my mind that the plane itself could
Be the primary instigator and greatly reduce external intruders.

I know this is all sci-fi now, but it's interesting to imagine the possibilities...

[bobbytables]

Vilnius - Automatics were engaged and were still available (autopilot A I believe would have worked) after B hydraulics were turned off. Would the aircraft have crashed anyway if the automatics were engaged? If you clear the warning system for the hydraulic failure and don't monitor flap deployment then even with the AutoPilot engaged, the outcome is uncertain.

We already have an example of a “more automated” aircraft than the B737. Absent at least one uncorrelated failure, an Airbus most likely would have saved the day in Vilnius.

[Zionstrat2]

Originally Posted by bobbytables

We already have an example of a “more automated” aircraft than the B737. Absent at least one uncorrelated failure, an Airbus most likely would have saved the day in Vilnius.

Yes you're absolutely right, the evolution of current processes and systems are "good enough".... Even with the recent events, overall safety is still better than it used to be.

The epiphany that hit me, is that safety needs to continue to get better and at some point and some time total automation will almost certainly become the next step.

I appreciate this thread reminding me that the day is probably well off in the future and we continue the pilot feeding the dog jokes.

[bobbytables]

Originally Posted by Zionstrat2

Yes you're absolutely right, the evolution of current processes and systems are "good enough".... Even with the recent events, overall safety is still better than it used to be.

The epiphany that hit me, is that safety needs to continue to get better and at some point and some time total automation will almost certainly become the next step.

I appreciate this thread reminding me that the day is probably well off in the future and we continue the pilot feeding the dog jokes.

a big chunk of recent safety advances has accrued to Airbus operators and not to Boeing operators because of Boeing’s desire (pushed by their airline customers) to keep commonality. It’s not a popular thing to say, but Boeing aircraft are less safe than Airbus aircraft. They demand more from the humans operating them, and humans are fallible.

[tdracer]

Originally Posted by bobbytables

a big chunk of recent safety advances has accrued to Airbus operators and not to Boeing operators because of Boeing’s desire (pushed by their airline customers) to keep commonality. It’s not a popular thing to say, but Boeing aircraft are less safe than Airbus aircraft. They demand more from the humans operating them, and humans are fallible.

If you except the 737MAX MCAS -up, the stats don't back that up. The 737NG and the A320 hull loss rates are nearly identical.

[EXDAC]

Originally Posted by bobbytables

It’s not a popular thing to say, but Boeing aircraft are less safe than Airbus aircraft. They demand more from the humans operating them, and humans are fallible.

Please reconcile those statements with the events of Air France flight 447.

[Zionstrat2]

Maybe save AB wars for other threads!)

But appreciate the thoughts on the future of total automation.

[Junkflyer]

Automation is fine. It does not always work the way is supposed to. If you have experience with it, it may be easier to understand that point.
What really matters is pilots reacting to and taking control when it does not go as planned. If pilots spend most of their time in the cockpit just pressing buttons, it leaves them less able to respond correctly when required.
Just my opinion, can't back it up with any statistics.
My company requires a hand flown single engine ILS approach and maybe go around during the yearly maneuvers validation checkride.. it is often dreaded, but a great way among. other problems to make sure your skills are where they should be.

[FullWings]

I think of the aviation automation issue in a very similar way to vehicles. As the automation level increases, you get to the point where the human has been so out of the loop that they can’t be expected to cope when suddenly handed back control. It’s why, IMO, we need to go straight from roughly where we are today to totally autonomous operation and not implement the intervening stages. The seeming push towards LH single-pilot ops is an example of this: if that one pilot happens to be in the toilet when an emergency happens, the aircraft must be able to cope with every conceivable malfunction without assistance. So why is the pilot there in the first place?

If pilots spend most of their time in the cockpit just pressing buttons, it leaves them less able to respond correctly when required.

And the issue is that because of that, the pilots of tomorrow will have much less experience to fall back on, as opposed to those of us who had a decade or two with very limited or no automation, so had to develop the skills needed. There will be much less opportunity to achieve familiarity and competence, such that you could successfully intervene in the event of automation failure, which may be subtle...

[Zionstrat2]

Originally Posted by FullWings

And the issue is that because of that, the pilots of tomorrow will have much less experience to fall back on, as opposed to those of us who had a decade or two with very limited or no automation, so had to develop the skills needed. There will be much less opportunity to achieve familiarity and competence, such that you could successfully intervene in the event of automation failure, which may be subtle...

Exactly.. this is what I meant when I said I've switched from thinking that pilots need a lot more basic training... (Of course in the short term that probably is still true.)

But in this future scenario we're discussing, the pilot will not have to have a lot of training because total automation should be able to deal with any unusual events automatically.

In an age of total automation, the pilot is essentially the whole frame parachute. They would not have a lot of day-to-day experience. everything they did would be in the simulator, and the only time they would be involved would be after a significant automation failure.. and yes, if that happened to be combined with extreme structural failures or other things this emergency pilot will not be a guru and yeah it might be over.

On the other hand if it's simply the failure of total automation, that emergency pilot should have the training to configure for landing, line up with a long runway and keep the nose pointed at the piano keys.

[framer]

For me it’s a hard subject to think about because I have no expertise in software or design. My inclination is to say “ no way, not for another 100 years” but then I realise I don’t really know what is possible.
For example, I can’t imagine how a fully automated aircraft would deal with GPS spoofing that results in an EGPWS Pull Up command on a nice day while on descent in a situation where the pilots would easily look out the window and know it shouldn’t be followed. Perhaps that’s an easy problem to solve but with my limited knowledge I can’t see it being done in the next 40 years.