Will computers replace pilots?

[Inverted25]

Well I did mention a single pilot Citation accident due to pilot incapacitation. So I'm pretty aware of the models that are. Don't you think there's a pretty good reason they're not certified for 135/121. Although there are plenty of single pilot 135 ops in Alaska and other small charter companies but they are all single engine or piston twins that don't carry a lot of pax. Also none of the single pilot jets go very far. I really wouldn't want to use a pee tube for 6-7 hours. Again, if I was an owner that's not typed or has zero flying experience, just sitting in the back, I'd want two bodies up there.

People have to remember as well the single pilot citation has existed since 1977. It's not like it's a new concept that has anything to do with modern computers or UAV

 

[PositionAndHold]

People have to remember as well the single pilot citation has existed since 1977. It's not like it's a new concept that has anything to do with modern computers or UAV

While I agree, my argument is this. We build in so much redundancy into modern part 25 airplanes why engineer the cheapest but most effective piece of redundancy out, another pilot? For an owner operator a SP Citation/Phenom can be a great tool. They're taking the risk with the people in back just like you would in 182. To have a pax sitting in the back that have zero aviation knowledge and a single pilot up front is just wrong. While it rare it can happen. Any know if Cape Air has ever had a pilot incapacitation accident?

 

[Roger Roger]

Until computers can drink booze and leer at waitresses half their age they'll never truly replace pilots.

 

[Roger Roger]

While I agree, my argument is this. We build in so much redundancy into modern part 25 airplanes why engineer the cheapest but most effective piece of redundancy out, another pilot? For an owner operator a SP Citation/Phenom can be a great tool. They're taking the risk with the people in back just like you would in 182. To have a pax sitting in the back that have zero aviation knowledge and a single pilot up front is just wrong. While it rare it can happen. Any know if Cape Air has ever had a pilot incapacitation accident?

There is suspicion that they did but it was never proven AFAIK and it was pilot only on a night repo flight.

 

[MikeD]

The problem again is that it really doesn't lead anywhere. One of the best attempts at what you are describing is in the Just Culture algorithm. In that, we recognize three categories, error, at-risk and reckless. Consider each of these in this context:

It leads us to many places, if followed in the matter in which I described it in my first post.

An error is considered an unintentional act. Can you avoid errors by just paying more attention? There is no evidence that this is true. While it might _appear_ that we can look at an event as say that the person was not paying attention, the truth is that when we reconstruct an event we are viewing it through hindsight and in doing so we can see a clear path from the outcome back to the actions. The path is much less clear when you look at it from the other way.

Not necessarily not paying attention. There are many errors that can be identified, whether through physical evidence, CVR/ATC recording, etc, from the pilot-forward point of view. Errors of misplaced priorities due to distraction(s), not doing the correct actions at the correct times, failing to fly the aircraft first and ending up hitting the ground, something attached to it or another aircraft when not intended. Etc, etc. There are many conclusions that can come from many different kinds of human actions and inactions, that learning can (and always should) be derived from in a FAR deeper manner than just "pay more attention".

At-risk is a condition where a person intentionally deviates from a procedure but for a good reason. We are not supposed to look at the outcome in this, but the fact is that when a person deviates as a normal course of business at it works out well we say "that person used good judgment", but when it turns out badly we castigate them. In the algorithm we look at this as an organizational issue, and if done right, we say that the procedure or policy is out of sync with reality and that needs to be fixed.

Again, there is alot of learning that can come from situations like these when it comes to figuring out whether so-called "best practices" are actually that, or are they really not. You have to peel back the proverbial onion in these cases and see why a pilot was doing a particular action. Was is the course of normal business, or was it an procedural deviation during an emergency, that ultimately went bad due to other contributory factors?

Here's an example of an accident that falls under this, and alot of procedural learning came from this one with regards to best practices and related risk factors, for dealing with a dead stick aircraft. Decisions were made with the best intentions during an inflight emergency, but a combination of secondary factors came into play to make what would've been a deviation on this particular day far more risky than had it been a CAVU day. That may not have been immediately apparent at the time, and the chosen course of action could not be reversed once it was committed to unfortunately. Nothing to castigate the pilot over, but things for the community of pilots who flew that same plane and similar ones, to seriously consider and figure out what they'll put into their bag of tricks if faced with a similar situation. As well as potenital for new procedural guidance to be generated if determined to be needed.

http://forums.jetcareers.com/threads/dead-stick-leads-to-a-dead-end.113095/#post-1560527

The third is reckless, and that is just those rare times someone just does something that they know to be wrong. That is outside of this discussion (although there is evidence that pilots of RPV's are more inclined towards this as they do not have personal "skin in the game").

And there are accidents of this type too, such as the Delta Air Lines pilot who decided to perform a loop over his family's farm in Texas in his F-16 while XC, and packed it into the farmhouse, morting himself.

RPV pilots actually do have skin in the game, in that, while they may not be onboard their aircraft, they are directly responsible for what their remotely piloted aircraft does, barring some technical situation completely beyond their control. So they may not have "physical" skin in the game, but they certainly have "liability" skin in the game. So for normal operations, if an RPV has an accident.....midair with a manned aircraft or impacts the ground and kills people on the ground, the pilot is just as responsible as if he'd been onboard the aircraft itself. In that respect, that would be their skin in the game. Same skin, just different location.

So, with that definition of "error", just saying someone made a mistake does little to help us, but that is precisely what we do with tools such as HFACS.

That's not what I advocate. As explained in my first post, that "error" term is merely a causal descriptor. Without expanding upon that and answering the many questions I posted that delve deeply into that error and defining what that error specifically was, then it's nothing more than just that: a descriptor. As I said, you won't find an accident where the probable cause is "pilot error. That is all. Case closed"; without there being any kind of expansion, investigation, analysis and discussion as to what that error was; so it can be understood to the best ability possible, and lessons derived from that disseminated as necessary.

 

[UAShopeful]

Well, the military is already forging that path. Aside from the many unmanned aircraft they already have, they have also converted an F-16 into an unmanned unit and have flown it successfully. Taking that one step further, look at the F-22. From what I understand, the cockpits are tiny, and the AF put very little effort into making sure the life support system was safe, with pilots going hypoxic at lower altitudes. Considering the airframe can withstand more G's than the human body is capable of, the weakest link in that platform is the pilot. If (and probably 'when') they convert that plane over to unmanned, it really will be the weapon it was designed to be.

As for civil aviation... I know everyone says that theres no way anyone would set foot in an unmanned 737, I say give it 10 years or less, and people will be all over them. If it drops their ticket price by 10$, they take it.

 

[UAShopeful]

Doesn't change anything, datacom issues alone would make that not probable for public transport.

Not necessarily. They can put the systems in place to keep datalink up. Depending on how you do it, it can get really costly, if satellites are involved, but it could be done without them. Check out this video. Note that the drones are going in and out through the stadium tunnels and no link issues. I'm guessing they have repeaters set up all throughout the stadium to prevent loss of datalink, but it's working for them...

 

[seagull]

Not necessarily not paying attention. There are many errors that can be identified, whether through physical evidence, CVR/ATC recording, etc, from the pilot-forward point of view. Errors of misplaced priorities due to distraction(s), not doing the correct actions at the correct times, failing to fly the aircraft first and ending up hitting the ground, something attached to it or another aircraft when not intended. Etc, etc. There are many conclusions that can come from many different kinds of human actions and inactions, that learning can (and always should) be derived from in a FAR deeper manner than just "pay more attention".

The issue is not that those are not happening but rather that it is incorrect to attribute that to the fault of the person who did the 'error'. They did not do these on purpose and, in fact, based on what we know about cognition, we can and _should_ expect that most people in the same situation would do the same thing. Finding these today tends to result in saying that the person who did them had the issue when in reality we should have anticipated them when we designed the system.

Again, there is alot of learning that can come from situations like these when it comes to figuring out whether so-called "best practices" are actually that, or are they really not. You have to peel back the proverbial onion in these cases and see why a pilot was doing a particular action. Was is the course of normal business, or was it an procedural deviation during an emergency, that ultimately went bad due to other contributory factors?

It is all too easy to see how the person should have noticed something in hindsight. If it was an intentional deviation (e.g. emergency) then it was not an "error", was it?

Here's an example of an accident that falls under this, and alot of procedural learning came from this one with regards to best practices and related risk factors, for dealing with a dead stick aircraft. Decisions were made with the best intentions during an inflight emergency, but a combination of secondary factors came into play to make what would've been a deviation on this particular day far more risky than had it been a CAVU day. That may not have been immediately apparent at the time, and the chosen course of action could not be reversed once it was committed to unfortunately. Nothing to castigate the pilot over, but things for the community of pilots who flew that same plane and similar ones, to seriously consider and figure out what they'll put into their bag of tricks if faced with a similar situation. As well as potenital for new procedural guidance to be generated if determined to be needed.

Same issue. Did a person do something wrong because they were trained inadequately (or even incorrectly) or was it a facet of cognition?

And there are accidents of this type too, such as the Delta Air Lines pilot who decided to perform a loop over his family's farm in Texas in his F-16 while XC, and packed it into the farmhouse, morting himself.

That is not an "error" but rather reckless. Whole different issue and different approach to fixing it.

RPV pilots actually do have skin in the game, in that, while they may not be onboard their aircraft, they are directly responsible for what their remotely piloted aircraft does, barring some technical situation completely beyond their control. So they may not have "physical" skin in the game, but they certainly have "liability" skin in the game. So for normal operations, if an RPV has an accident.....midair with a manned aircraft or impacts the ground and kills people on the ground, the pilot is just as responsible as if he'd been onboard the aircraft itself. In that respect, that would be their skin in the game. Same skin, just different location.

However, they are not actually in the vehicle and thus cannot sense some things that are very challenging to notice. A slight change in sound, a vibration or odor come to mind as things that pilots have noticed and taken immediate action and thus saved an aircraft, all well before the system itself (even with very sophisticated warning systems) was aware of an anomaly.

In addition, there are numerous studies that show that people will take risks that they would not otherwise do, liability or not. Would the plot of such a vehicle even be liable? Under a just-culture atmosphere they really would not be unless it was a reckless act.

That's not what I advocate. As explained in my first post, that "error" term is merely a causal descriptor. Without expanding upon that and answering the many questions I posted that delve deeply into that error and defining what that error specifically was, then it's nothing more than just that: a descriptor. As I said, you won't find an accident where the probable cause is "pilot error. That is all. Case closed"; without there being any kind of expansion, investigation, analysis and discussion as to what that error was; so it can be understood to the best ability possible, and lessons derived from that disseminated as necessary.

NTSB just came out with a video highlighting a number of accidents which they described as "procedural non-compliance" as if that explained anything....

 

[seagull]

Well, the military is already forging that path. Aside from the many unmanned aircraft they already have, they have also converted an F-16 into an unmanned unit and have flown it successfully. Taking that one step further, look at the F-22. From what I understand, the cockpits are tiny, and the AF put very little effort into making sure the life support system was safe, with pilots going hypoxic at lower altitudes. Considering the airframe can withstand more G's than the human body is capable of, the weakest link in that platform is the pilot. If (and probably 'when') they convert that plane over to unmanned, it really will be the weapon it was designed to be.

As for civil aviation... I know everyone says that theres no way anyone would set foot in an unmanned 737, I say give it 10 years or less, and people will be all over them. If it drops their ticket price by 10$, they take it.

Military is a specific application where removing the pilot _reduces_ the risk to human life (at least for those that would otherwise be in the vehicle).

 

[seagull]

Not necessarily. They can put the systems in place to keep datalink up. Depending on how you do it, it can get really costly, if satellites are involved, but it could be done without them. Check out this video. Note that the drones are going in and out through the stadium tunnels and no link issues. I'm guessing they have repeaters set up all throughout the stadium to prevent loss of datalink, but it's working for them...

You need to get to 10^-7 or so reliability and I know of no wireless data that is close to that. In fact, I am not sure we are there for wired data.