When will the public accept full automation

Rotor dude was right. This is a troll thread....

Who starts a thread and never comes back?

Brien started this one and the little ***** hasn't been back since. Troll fvcker.
 
denverpilot said:
They're also ignoring how many times their computer blew up from bad code. LOL. A common mistake in computing and automation.
Click to expand...
If done right....the system should be fault tolerant....and continue with the last data upload.
 
denverpilot said:
Not only that but you have an encryption and repudiation issue. You're going to need a really secure data channel to pull off the ground command thing.

Or we can be stupid and start with unencrypted, spoofable links. And learn that lesson the hard way.
Click to expand...
Meh.....they'll get that figured out. The military did....with TCDL.:D
 
Dav8or said:
Care to enumerate these failure modes? If you can so easily identify them, could you not just as easily come up with solutions? Why is flying an airplane from point to point too complicated for a computer to accomplish?

In the case of AF477, the pilots were in control and crashed. How could the computer do any worse? In that case, people say that nobody was flying the airplane. Could this not be a case where the computer steps in and flies the airplane when no one else is?

I was suggesting a symbiotic relationship. If the computer is screwing up due to software glitch, or bad data from a sensor, such as a pitot tube, the pilots would have the authority to identify, over ride and correct this problem. In addition, my proposal was to leave cockpit management and authority just as it is now with humans in control.

The HAL 9000 only takes over when it can identify stupid pilot behavior, like directing the airplane to fly into the ground, or using the controls to enter a spin, or flying over the ocean and pushing the button to dump fuel. It could also assist when the pilots are busy on their lap tops and fly 100 miles past their airport, or the pilots are making a terrible approach to an airport.

I'm sure I could be completely wrong and such a system will never be built due to the complexities and potential dangers, but I doubt it. If I thought of it, that means there are people already working on it.
Click to expand...

AF447 was initiated by an instrumentation failure. The automation failed and the crew was not smart enough to compensate.

All automatic systems can suffer input problems. And hell no, I won't enumerate them all. That is EXACTLY the problem. Leave one out and everybody dies.

You're obviously not a systems engineer, as you don't seem to realize you've asked me to do several months of analysis for you. And Systems Engineering Question #1 is ALWAYS "Is this worth doing?" In this case it's obviously a very complex and risky solution to an exceedingly rare problem.

One failure mode is that the pitot tubes ice, so the computer thinks the pilot is trying to rip the wings off. Another is an emergency descent. What about disabled systems? Is it suicide to disable safety equipment because, say, it might be smoking or malfunctioning?

Truly half baked idea. I'm finding it hard to believe you're serious. You didn't post it on April 1,
 
Last edited:
MAKG1 said:
AF447 was initiated by an instrumentation failure. The automation failed and the crew was not smart enough to compensate.

All automatic systems can suffer input problems. And hell no, I won't enumerate them all. That is EXACTLY the problem. Leave one out and everybody dies.

You're obviously not a systems engineer, as you don't seem to realize you've asked me to do several months of analysis for you. And Systems Engineering Question #1 is ALWAYS "Is this worth doing?" In this case it's obviously a very complex and risky solution to an exceedingly rare problem.

One failure mode is that the pitot tubes ice, so the computer thinks the pilot is trying to rip the wings off. Another is an emergency descent. What about disabled systems? Is it suicide to disable safety equipment because, say, it might be smoking or malfunctioning?

Truly half baked idea. I'm finding it hard to believe you're serious. You didn't post it on April 1,
Click to expand...

and when thinking about failure modes, don't forget combination of events.

And one of the most difficult things to account for in a failure analysis are all the degraded modes (e.g., partial failures).
 
Checkout_my_Six said:
No worries.....this will be a slow transition.
Click to expand...
...and it's already happening.

TLDR: Just because Angry Birds crashes your iPad doesn't mean all of your flight control computers will dump and lock up at the same time.

Seems like a lot of the discussion here is focused on up/downlink, and that's reasonable, but there also seems to be a general mistrust of flight-critical computers and software. There are *already* airplanes flying with control systems and flight control computers that have the *authority* (meaning control available, not permission) to pull the wings off an airplane or cause it to depart controlled flight [faster than a pilot can react - nauga] Designers and testers spend thousands (and thousands and...) of hours designing, testing, and analyzing these systems before widespread release. Much of this focuses on failures, faults, reversions, and degraded modes to make sure this doesn't happen.

They're flying *now* and have been with regularity since the late '70's. Do things slip through the cracks? Certainly. Are these post-test escapes catastrophic? None that I can recall. What's the leading mishap cause in airplanes with these types of control? Pilot/Aircrew error.

I think there are technical hurdles (like detect and avoid) that remain, and cost and time of development and integration and public acceptance will keep full autonomy out of commercial cockpits for a long time, but the technology that many seem to fear is already here and flying.

Nauga,
and his steel toe reboot
 
Last edited:
mkosmo said:
The NASA SPO project? We've been working on it in our office, too. You should read some on it. Not quite ground command for what we're working with, but more of the super dispatcher position.
Click to expand...

Totally separate , was born in the 747-8 program and initiated somewhat by Boeing , more of a tech upgrade for avionics
 
So, nauga, which of these systems operating now REMOVE authority from pilots without possibility of override, based on a fuzzy and poorly defined decision tree?

Most of the reversions give authority to the pilot in order to deal with a failure.

Such a system is probably not impossible with infinite budget, but it is a very foolish strategy to blow that kind of money for the fear du jour.

There is little question these days that the failure modes of multiple automation systems as well as serious interface errors played a significant role in the AF447 crash. While it clearly is not exclusive fault, accidents like that seldom have one cause. So it's a little misleading to say there haven't been fatalities.
 
Last edited:
MAKG1 said:
So, nauga, which of these systems operating now REMOVE authority from pilots without possibility of override...
Click to expand...
The automatic ejection system in the Yak-141 :rolleyes: It doesn't (didn't) just remove authority, it removed the pilot. :D

If you think I'm condoning a reduction of pilot authority in degraded modes you've read way too much into my post.

Nauga,
who thinks the future will be better tomorrow
 
So....is it time to start talking data-bus architecture and primary and secondary data links? :D
 
taters said:
Totally separate , was born in the 747-8 program and initiated somewhat by Boeing , more of a tech upgrade for avionics
Click to expand...
Interesting. Do you know of any published information on the project? I'd be curious to read it.
 
MikeELP said:
https://www.youtube.com/watch?v=qDrDUmuUBTo


:D
Click to expand...


Did you ever read 2010? Clarke wrote a pretty good follow on that explained how the idiot humans had made HAL9000 ****zophrenic and created a logic problem that could only be solved by killing the human crew. It was a pretty good read.

The movie with Roy Scheider was changed fairly significantly from the book and was entertaining also, but if you hadn't read the book, they left some stuff out that kinda left movie-only audiences not understanding quite what happened. They completely removed all but a single speech from Dr. Chandra on HAL's neuroses and Clarke did a better job explaining it in the book.

Fun stuff. One of the more interesting lines in the book is when two Astronauts are talking about what they miss back home and one mentions the color green. It doesn't exist much anywhere else in our solar system other than Earth in any significant quantity.
 
It's wildly premature to talk about architectures when you don't even have a sensible and complete use case. Agile techniques are not at all appropriate to safety critical systems.

If the use case is wildly stupid -- and it is in this case -- even the best architecture in existence will still produce a stupid system.

You gotta figure out what you're doing before figuring out how.

Operating with the last received message is not always sensible. Sometimes, perhaps even often, it makes more sense to flip an annunciator and drop to a lower level of automation.

I can tell you guys don't do much with controls.
 
Last edited:
MAKG1 said:
It's wildly premature to talk about architectures when you don't even have a sensible and complete use case. Agile techniques are not at all appropriate to safety critical systems.

If the use case is wildly stupid -- and it is in this case -- even the best architecture in existence will still produce a stupid system.

You gotta figure out what you're doing before figuring out how.
Click to expand...

are we ready to go to fly by wire systems?....or are we still talking about cables and hydraulic assist?
 
MAKG1 said:
FBWs are here, but it is debatable whether the added complexity really is such a great idea. But that has little to do with anything at hand.
Click to expand...

Oh, quite contrare.....FBW systems are completely digital and ripe for datalink integration.

In fact the newer systems are all setup.....and almost ready to go. :D
 
Checkout_my_Six said:
Oh, quite contrare.....FBW systems are completely digital and ripe for datalink integration.

In fact the newer systems are all setup.....and almost ready to go. :D
Click to expand...

Remote control with passengers may be possible or even easy, but it's still a stupid idea.

It's a huge engineering mistake to confuse ease with sensibility. You're skipping the first, most important, step.

This is one of those cases where you can work out the nominal behaviors easily. Nominal behaviors seldom kill. It's the off nominals where the problem lies. Good luck enumerating them.

And, just what problem are you trying to solve? One lone nutcase in 20 years? Nut cases can be on the ground, too, and with more critical pieces, they have more opportunities, and a much larger physical security problem.
 
Last edited:
There's an interesting series of books by Alvin Toffler that you might read called The Third Wave(1980). One of the themes of the book is that the limit of technological advancement is man's willingness to embraced change. Old peple might have cell phones but very few of them are using them to their abilities.

I program computers. I will never fly in an airplane controlled solely by one.
 
MAKG1 said:
Remote control with passengers may be possible or even easy, but it's still a stupid idea.

It's a huge engineering mistake to confuse ease with sensibility. You're skipping the first, most important, step.

This is one of those cases where you can work out the nominal behaviors easily. Nominal behaviors seldom kill. It's the off nominals where the problem lies. Good luck enumerating them.

And, just what problem are you trying to solve? One lone nutcase in 20 years? Nut cases can be on the ground, too, and with more critical pieces, they have more opportunities, and a much larger physical security problem.
Click to expand...

you're right....on second thought, it'll never work. :rofl:
 
too often people don't consider the new failure modes caused by technical solutions intended to address old failure modes.
 
denverpilot said:
Self driving cars are going to be awesome. You can leave instructions in your Will that the car will be programmed to go all the places you never got a chance to go when alive, and that until your money runs out, it'll pay for someone to arrange to have your car programmed to go to all those places and be refueled along the way, so the dead you can ride around, see the sights, and decompose the car. ;)

Hahahaha.

Okay not as unlikely: we will see news reports of a self diving car or fully automated aircraft showing up somewhere with a dead guy in it.

They'll probably then require that the aircraft or car check our pulse through our asses or something and re-route to the closest emergency services location if we have croaked so cars and airplanes don't show up at loved one's houses with dead relatives in them or airports in equipped to handle a dead guy.

Heh. Twisted ain't it?
Click to expand...

Twisted? Perhaps, but a good engineer always thinks about what could go wrong.
 
txflyer said:
Rotor dude was right. This is a troll thread....

Who starts a thread and never comes back?

Brien started this one and the little ***** hasn't been back since. Troll fvcker.
Click to expand...

I don't see a problem. People are obviously interested in the topic, and as far as I know, participation in the thread is voluntary.
 
cowman said:
I don't necessarily have a problem with it if there's a human who can take over. It could be a one pilot crew and he could just nap or read a book on the flight over but I think you'd need that before anyone would be ok with it.
Click to expand...

How would that pilot maintain proficiency?
 
Bob Noel said:
too often people don't consider the new failure modes caused by technical solutions intended to address old failure modes.
Click to expand...
Certification of flight critical software *requires* such consideration.

Nauga,
at the validation station
 
Palmpilot said:
I don't see a problem. People are obviously interested in the topic, and as far as I know, participation in the thread is voluntary.
Click to expand...

Yep right on, there always are the little people who think badly of others, little people little minds, who judge other by them selves.
 
The technology to build such a system is very simple. It's just a basic supervised machine learning classification system, the algorithms are not that complicated and they are very well established. The hard part is generating all the training data; you'll need to have a lot of pilots flying a lot of time in simulators, and even that might not realistically generate enough data to train the system such that it has adequate precision/recall (catches the bad intentioned pilots reliably, doesn't label the good intentioned pilots as bad). That being said, you can ensure far better accuracy for one if you sacrifice the other; the system could be tuned to only turn on in cases where pilot behavior was so egregious, there was no other explanation than bad motive. For instance, a pilot pulling circuit breakers and disabling key monitoring systems for no reason.

The system may also work well as a means to preemptively alert people on the ground of abnormal behavior, and with some sort of integrated override system, you could allow someone on the ground to quickly assess and take over the plane's flight controls as necessary.
 
Last edited:
Tom-D said:
Yep right on, there always are the little people who think badly of others, little people little minds, who judge other by them selves.
Click to expand...


And there are people who let other people spit all over them and say 'thank you, may I have another?'

Be honest. Have you ever STARTED a thread and just abandoned it?
 
txflyer said:
And there are people who let other people spit all over them and say 'thank you, may I have another?'

Be honest. Have you ever STARTED a thread and just abandoned it?
Click to expand...

That's funny, I've checked all over myself, and I can't find any spit anywhere.

Why do you equate that with being spat on? :confused:

Basically, I just can't think of any reason to care what other people's reasons are for starting threads, and even less reason to care what other people think others' reasons are. I have no emotional investment whatsoever in whether a person who starts a thread participates in the discussion. The ONLY thing I care about when I look at a thread is whether there is an interesting discussion going on.
 
Last edited:
Palmpilot said:
That's funny, I've checked all over myself, and I can't find any spit anywhere.

Why do you equate that with being spat on? :confused:

Basically, I just can't think of any reason to care what other people's reasons are for starting threads, and even less reason to care what other people think others' reasons are. I have no emotional investment whatsoever in whether a person who starts a thread participates in the discussion. The ONLY thing I care about when I look at a thread is whether there is an interesting discussion going on.
Click to expand...


That's cool, but rotor dude pointed out don't feed the troll and I watched and waited and here we are eight pages later and the OP hasn't been back since... don't you find that a tad bit strange?

I'll ask you to be honest too. Have you ever started a thread and never looked at it again?

I could care less, automation is a hot topic obviously. I was just pointing out the queerness of starting a thread and then vanishing. I hope the dude isn't dead or worse ... :dunno:.
 
txflyer said:
That's cool, but rotor dude pointed out don't feed the troll and I watched and waited and here we are eight pages later and the OP hasn't been back since... don't you find that a tad bit strange?
Click to expand...

No, but why does it matter?

I'll ask you to be honest too. Have you ever started a thread and never looked at it again?
Click to expand...

Probably not, but we don't know whether he has looked at the thread again or not, and I still can't think of a reason to care.

I could care less, automation is a hot topic obviously. I was just pointing out the queerness of starting a thread and then vanishing. I hope the dude isn't dead or worse ... :dunno:.
Click to expand...

Since you've switched from calling him names to being concerned for his well being, I can ease your mind on that point. If you check the statistics section of his user profile, you will see that he has accessed the forum as recently as this afternoon.
 
Palmpilot said:
No, but why does it matter?

Seems odd to me .. :dunno:

Probably not, but we don't know whether he has looked at the thread again or not, and I still can't think of a reason to care.

Good for you.

Since you've switched from calling him names to being concerned for his well being, I can ease your mind on that point. If you check the statistics section of his user profile, you will see that he has accessed the forum as recently as this afternoon.
Click to expand...

That's good. I was starting to worry about the troll fvcker... :lol:
 
txflyer said:
Be honest. Have you ever STARTED a thread and just abandoned it?
Click to expand...

Yes, I often ask questions and let the thread go where it will just to see others opinions.
 
MAKG1 said:
AF447 was initiated by an instrumentation failure. The automation failed and the crew was not smart enough to compensate.

All automatic systems can suffer input problems. And hell no, I won't enumerate them all. That is EXACTLY the problem. Leave one out and everybody dies.
Click to expand...

What would a "smart" crew have done? When I did instrument training, they taught me to test a questionable reading on an instrument by cross checking with other instruments that get their data from other inputs. Can a computer not be instructed to do the same, or just you can't imagine that possibility?

You're obviously not a systems engineer, as you don't seem to realize you've asked me to do several months of analysis for you. And Systems Engineering Question #1 is ALWAYS "Is this worth doing?" In this case it's obviously a very complex and risky solution to an exceedingly rare problem.

One failure mode is that the pitot tubes ice, so the computer thinks the pilot is trying to rip the wings off. Another is an emergency descent. What about disabled systems? Is it suicide to disable safety equipment because, say, it might be smoking or malfunctioning?
Click to expand...
And you're obviously not an engineer. You sound more like a corporate numbers cruncher. An engineer is tasked with , or identifies a problem and then creates a solution. They don't just give you 85 reason's not to even try. If it were up to you, the "systems engineer", the powered airplane as we know it, would never have flown.

It's true that this anomaly is exceedingly rare, but if the people demand it, the people will get it. 9/11 ring any bells here??

Truly half baked idea. I'm finding it hard to believe you're serious. You didn't post it on April 1,
Click to expand...
I find it hard to believe you call yourself an engineer. A truly half baked attitude it would seem. Seriously, open your mind and take on challenges. I guess it's the difference between middle management and the history books. That's OK. Everybody needs a job and an income, I get it.
 
Last edited:
Any one here ever heard of GinGO?

That's what you get when two engineers are trying to get a computer geek to write code, from an idea.
 
It will start with cargo flying the oceans (not over populations).
The planes will be semi automated much like the drone missions in the mideast today. A 'pilot' somewhere in the world monitoring the progress and taking over if something beyond the automation parameters occurs. This will be sooner rather than later.

Now passenger flights - gonna be awhile before that happens.
 
You must log in or register to reply here.
Back
Top