Airbus A320 Down

Riiiiiight.........
If it's made by a human and or maintained by a human there's always going to be a level of imperfection. Humans - we're almost like a virus.

As I understand it, a new SkyNet will be developed to control all airline flights in Europe. The proposed SkyNet will operate on the heuristic programmed algorithmic computer version 9000 series computers in redundant, ultra-secure sites located at multiple points on the continent and in the UK. The natural expansion of the system will be to control all airline flights worldwide thus preventing all future airline disasters. The plan is pure genius and foolproof at all levels. It will also provide socialist employment of hundreds of thousands of people through direct and indirect economic stimulation.
 
Doubtful. The only people that will fly on autonomous aircraft are the tech freaks who think driverless cars are awesome.

Maybe, not the way I see it, and industry would love to get rid of pilots, we're an expensive PITA especially when when they already pay for the automation anyway. One year of industry pilot wages pays for the conversion. Humans have a bad record in the cockpit, that's a fact, automation has the statistical advantage.

Anything can be sold to the consumer public through the use of fear, and autonomous airplanes are an easy sell from the non aviator consumer point of view. Look at all the accidents over the last 2 decades, human factor crashes to human factor saves (that could be replicated with automation, no worries) are 10:1 or greater.

You are projecting your opinion as a pilot on the consumer public, and it is not so, people think flying is oh so difficult, and they are now all carrying iPhones and are comfortable living their lives remotely via technology. They will be far more accepting of autonomous airliners than you think. Eventually even grandma will get on one if that's what she has to do to see her grand child.

Actually, you should look forward to this, because it will also lock step with Amazon's autonomous delivery drone program that has finally been green lighted by the FAA for development. This will revolutionize small aviation transportation as well, as Amazon has the capital resources and commitment to develop the propulsion technology that will make carrying humans safely around town, from the garage to the store, work, or grandmas house direct in a "Directed Autonomy" mode operating it like you were using Siri, "Good morning Alice, take me to work please, wake me when we get there." "I'll have you there in 15 minutes, have a nice nap."

You have to realize this has been in the works for decades, it is what NextGen has always been about, directed autonomous flight. We have been piecing together all the disparate technological systems, and we are getting to the end of the game. Bringing the Amazon capital resources into play at the mid-low weight level of the game will solve the issue that has always plagued 'flying cars', and that is a viable power supply. Now if Amazon gets together with LockMart Skunk Works and supercharges the CFR program, holy crap we could have something.
 
Maybe, not the way I see it, and industry would love to get rid of pilots, we're an expensive PITA especially when when they already pay for the automation anyway. One year of industry pilot wages pays for the conversion. Humans have a bad record in the cockpit, that's a fact, automation has the statistical advantage.

Anything can be sold to the consumer public through the use of fear, and autonomous airplanes are an easy sell from the non aviator consumer point of view. Look at all the accidents over the last 2 decades, human factor crashes to human factor saves (that could be replicated with automation, no worries) are 10:1 or greater.

You are projecting your opinion as a pilot on the consumer public, and it is not so, people think flying is oh so difficult, and they are now all carrying iPhones and are comfortable living their lives remotely via technology. They will be far more accepting of autonomous airliners than you think. Eventually even grandma will get on one if that's what she has to do to see her grand child.

Actually, you should look forward to this, because it will also lock step with Amazon's autonomous delivery drone program that has finally been green lighted by the FAA for development. This will revolutionize small aviation transportation as well, as Amazon has the capital resources and commitment to develop the propulsion technology that will make carrying humans safely around town, from the garage to the store, work, or grandmas house direct in a "Directed Autonomy" mode operating it like you were using Siri, "Good morning Alice, take me to work please, wake me when we get there." "I'll have you there in 15 minutes, have a nice nap."

You have to realize this has been in the works for decades, it is what NextGen has always been about, directed autonomous flight. We have been piecing together all the disparate technological systems, and we are getting to the end of the game. Bringing the Amazon capital resources into play at the mid-low weight level of the game will solve the issue that has always plagued 'flying cars', and that is a viable power supply. Now if Amazon gets together with LockMart Skunk Works and supercharges the CFR program, holy crap we could have something.

lol, "NextGen" is in no way capable of the above. Could the above be done? Yes. But you'd need to knock the FAA out of the equation with the coordination of all those flights.

Getting it done on a national level would be an incredibly impressive technological feat and given all the obstacles I'm not sure I'll see it in my life.
 
"Maybe someday AI pilots will be perfected enough to completely take over the cockpit without a back-up plan, but that day is not in our foreseeable future," Lin tells io9. "We can't even build much simpler devices, such as laptop computers and smartphones, that never crash in much more ordinary conditions."

http://io9.com/a-i-pilots-are-not-the-solution-to-preventing-airline-1694054826/+AnnaleeNewitz

I tend to agree with the article. I think the days of completely autonomous AI flight are far in the future. Current computer technology is too prone to failure.

Anyone who's accidentally coded an infinite loop knows what I'm talking about. :yikes:
 
lol, "NextGen" is in no way capable of the above. Could the above be done? Yes. But you'd need to knock the FAA out of the equation with the coordination of all those flights.

Getting it done on a national level would be an incredibly impressive technological feat and given all the obstacles I'm not sure I'll see it in my life.

NextGen is a development program, not a specific resouce. It is an operating model. The program has already antiquated the initial technology, but that technology can still provide the service to make the system function, because the requirements for function haven't changed. While it isn't current generation technology, it is time tested and has gone through several generations of improvement even before it's full deployment date, and will continue to do so. There's no point in upgrading from the technology we have used to grow into a cohesive system until that initial cohesive system has been implemented and and gone through the initial v1.0 version in completed form and shook it down. Right now all we are doing is the Beta on the components of the system. If you want to know where this is headed, look through science fiction, it's the proven best predictor, they just usually have the timeline wrong.
 
Take a look at the spec of the Lockeed 1011. It would fly itself to a landing. That was years ago.

I have trouble seeing why pilots think they are so special their job can't be automated? It's delusional in fact in the light of everything that history tells us from the Civil War on. Anything to take the cost of maintaining humans out of the spread sheet happens, period. We already run military drone programs under directed autonomy. Military pilots jobs are already being supplanted by autonomy. Wake up pilots, your cushy career may be about to come to an end.
 
I have trouble seeing why pilots think they are so special their job can't be automated? It's delusional in fact in the light of everything that history tells us from the Civil War on. Anything to take the cost of maintaining humans out of the spread sheet happens, period. We already run military drone programs under directed autonomy. Military pilots jobs are already being supplanted by autonomy. Wake up pilots, your cushy career may be about to come to an end.

AI baby, a friend who works in this field says 25 to 30 years max, most jobs, like pilots, drivers, factory control, financial, cpa and other functions will be done by machines due to pending advances in artificial intelligence.
 
AI baby, a friend who works in this field says 25 to 30 years max, most jobs, like pilots, drivers, factory control, financial, cpa and other functions will be done by machines due to pending advances in artificial intelligence.


As a CPA, I'm not sure I want Hal doing my return. ;)
 
I love computers, I love automation... I used to be a programmer. Computers are great at doing simple predictable tasks with great precision, speed, and lack of errors. However, they're still really really bad at decision making. I will grant you it is getting pretty good these days but they have no ability to go beyond scenarios/parameters laid out by the engineers and programmers who designed them.

I don't even mind the idea of the computer driving or flying, they are less likely to make a mistake. However, I still want a human in the loop somewhere in case the machine does something stupid or creative solutions are called for.
 
AI baby, a friend who works in this field says 25 to 30 years max, most jobs, like pilots, drivers, factory control, financial, cpa and other functions will be done by machines due to pending advances in artificial intelligence.

This event could catalyst the investment to speed that time table. The development is sufficient, primarily just implementation and deployment costs, and if industry gets a green light, they will throw all the money required at it to make it happen ASAP, because it will cost the same at the end to implement, but they start saving the day they get the first directed autonomous on line service.
 
...why is there still manned control when the plane already auto lands when the pilots can't anymore...
How about when the plane can't autoland anymore? I don't know about the newest 787, but here are the crosswind landing limits for a B777F:

Landing: 38kts
Autoland: 25 knots
Cat I/II Autoland: 15 knots
Autoland headwind: 25 knots

Above 25 knots of cross- or headwind, it's a manual landing. Period. Can they design it for more. I'm sure, but at what cost?

...we're an expensive PITA especially when when they already pay for the automation anyway. One year of industry pilot wages pays for the conversion. Humans have a bad record in the cockpit, that's a fact, automation has the statistical advantage.
Are you kidding me? We are extremely CHEAP compared to what they'll have to do to replace us.

Look at all the accidents over the last 2 decades, human factor crashes to human factor saves (that could be replicated with automation, no worries) are 10:1 or greater.
Really? I thought you were smarter than that. You have no "fully automated crashes/saves" data to compare this to. Garbage argument.

You are projecting your opinion as a pilot on the consumer public, and it is not so, people think flying is oh so difficult
No. It's not. We all know that. Decision making and cockpit management during abnormal operations is what's difficult.

I have trouble seeing why pilots think they are so special their job can't be automated? ... We already run military drone programs under directed autonomy. Military pilots jobs are already being supplanted by autonomy. Wake up pilots, your cushy career may be about to come to an end.
If you're hanging your hat on the military RPA program, you had better find another argument.

I'm so mad that you made me actually get up, bust out my laptop to reply to this.

I've only been on this board a short time, but I'm starting to figure out some of the "personalities" here. I get you, Henning.

You talk about autoland. If you've ever been on the flight deck of an airliner that's shooting a Cat III, it's a quiet, tense environment. We're not doing anything but monitoring, but knowing that if the AFDS reverts to LAND2 or NO AUTOLAND at 100 feet, we're going to be busy. And it happens... trust me.

I have friends that are RPA pilots in the military. Their safety record is abysmal. They crash drones at an outstanding rate. And we're not talking combat losses. We're talking comm losses. Datalink goes down, spikes in comms, sunspots... whatever. They crash, a lot. Add to that that the infrastructure to launch and recover these RPAs is about twice that of a manned flying squadron. The comm equipment alone would be an astronomical cost to upgrade.

With the F-35, we have seen the last of the manned fighters. About 1/3 of the weight of a fighter is there just to support the fuzzy, pink body sitting in it. Pressurization system, seat, avionics... all that gets eliminated when you take the pilot out. For a passenger airplane, all that needs to be there. You need to pressurize it for the passengers, you need toilets, water, seats. The pilots are just an add-on which amounts to a rounding error.

Next, you'll have to build a whole new fleet of aircraft. From 737 size to 747 size. I don't think retrofitting the current fleets is going to work. Flying a ETOPS flight, there is so much that goes into it as far as redundancy in the systems. Hydraulic systems... redundant. Electrical systems... redundant. But we pilots are still there when the redundancy fails. Generator failed. Dispatch it with an MEL. Okay, now we're down to one. If we lose the other one, the pilots are still there that can hand fly it down. No pilots, you better figure out a couple more redundancies.

QRH says to pull and reset a circuit breaker. Okay, I'll reach up there and pull and reset it. What if I'm not there. We'll have a servo built in to do it. But make sure each circuit breaker has three servos. One to dispatch on an MEL, and one redundant one in case the second one fails. We pilots are cheap redundancy.

Also, plan on the expense of upgrading datalink communications world-wide. There are lots of places we fly that we can't get datalink service... heck, we can barely get HF communication. You better be prepared to pay to upgrade/build world-wide triple-redundant, hack proof data-link communication that covers 90% of the globe from Nigeria to Mongolia. And you better find the satellite band-width to do it.

Add to that the cost of getting Cat III ILSs to every commercial airport that would be served by a RPA. As you know, there are different quality of ILS signals. Only certain ILSs are certified for full-up Cat III autolandings. Make sure you have the funds to upgrade every ILS out there to the Cat III standard.

The other reason that the military can run their RPA program is that they launch, what? 150-ish RPA sorties a day? How about an airline like American that are launching 3,000 flights in a day over 350 destinations worldwide. Good luck with that.

I know my job is replaceable. I know that technology might be able to replace me, but, in my mind, there are such huge barriers to accomplish this, I don't think I'll see it in my lifetime.

A pilot goes nuts, locks the captain out and crashed an airplane. Now we have cries of "get the pilots out of there!" What happens when a fully automated airliner crashes (and it would happen), and then we'll hear cries of "a pilot would have been able to save that! Put pilots back in!"

All these "skynet is self-aware" guys need to ask themselves why we don't have self-driving locomotives carrying all our passengers and freight?

Henning, you're a boat guy... Why aren't there unmanned cruise ships sailing to the Bahamas?

Those are much simpler machines, operating in much more predictable environments. The logic that computers are gonna take over all passenger transportation-just because computers are getting "smart" is flawed. If that were the case, locomotives would definitely have become fully automated at least 50 years ago.

The type of intelligence required to operate airliners in complex environments is something computers fall far short of. They just can't handle approximations, novel situations, and analogy well at all. Not even on the level of a small child.

I am not worried at all about that threat to our careers. There are many much bigger fish to fry.
 
Last edited:
I love computers, I love automation... I used to be a programmer. Computers are great at doing simple predictable tasks with great precision, speed, and lack of errors. However, they're still really really bad at decision making. I will grant you it is getting pretty good these days but they have no ability to go beyond scenarios/parameters laid out by the engineers and programmers who designed them.

I don't even mind the idea of the computer driving or flying, they are less likely to make a mistake. However, I still want a human in the loop somewhere in case the machine does something stupid or creative solutions are called for.

There are very few decisions that need to be made, and humans fail at them more often than not.:dunno: Flying an airliner is not particularly subjective, and with the terrain database the Space Shuttle program developed, the computer is much more capable of finding the best landing situation in any given range than a human pilot in an emergency. An autonomous aircraft would have landed Sully's flight at JFK, IIRC the plane would have made it safely. Sully chose the water (correctly IMO) because by his human limitations on being able to calculate definitively being able to make a runway, he chose the water, a situation less than optimal, but giving a much higher probability of survival with no collateral losses compared to failing to make the runway. An autonomous system is not limited in the same manner. Also look at AF447, three human safety links all failed at decission making. Humans will not be tolerated to be in direct control of airliners for much longer.
 
Last edited:
I have trouble seeing why pilots think they are so special their job can't be automated? It's delusional in fact in the light of everything that history tells us from the Civil War on. Anything to take the cost of maintaining humans out of the spread sheet happens, period. We already run military drone programs under directed autonomy. Military pilots jobs are already being supplanted by autonomy. Wake up pilots, your cushy career may be about to come to an end.

What about boat captains ? :D
 
AI baby, a friend who works in this field says 25 to 30 years max, most jobs, like pilots, drivers, factory control, financial, cpa and other functions will be done by machines due to pending advances in artificial intelligence.

And due to the desire to cut costs.

BTW, the company leading the way in AI is Google - it's no secret that they seek to develop and implement AI. One of the founders fathers was involved in AI research.

Think of that every time you let them place a cookie, scan your email, or track your android phone everywhere you go.
 
There are very few decisions that need to be made, and humans fail at them more often than not.:dunno: Flying an airliner is not particularly subjective,.

what a crock of bull****e

humans are a system of the airplane and always will be ,all the technological and automated systems are ultimately human designed , coded , operated ..and therefore they can FAIL

last night going into ATL on the KOLLT arrival in the pitch black and being vectored around some storms ..we caught a piece of one ( not even represented on radar ) and got into strong moderate turbulence which required immediate evasive action and the "sitting down " of the flight attendants..A cascade of command decisions took place that I could never imagine a computer making. Computers aren't decision makers , they are data processors ie garbage in = garbage out
 
Look, you can live in denial, or prepare for the future. If you need to have good work 20 years from now, you need to prepare for an alternate income stream besides Air Carrier pilot, be it pax of freight. The future is coming whater you want it to or not. May as well make sure it's going to be something you like.
 
Look, you can live in denial, or prepare for the future. If you need to have good work 20 years from now, you need to prepare for an alternate income stream besides Air Carrier pilot, be it pax of freight. The future is coming whater you want it to or not. May as well make sure it's going to be something you like.
Henning, it is you who are in denial. Did you even read my post #614? Any answers as to the reality of implementing this system? To the cost and logistics involved? Could you get an autonomous aircraft to takeoff from Point A and fly to Point B in benign conditions, sure. But to replicate that for 100,000 flight each day across the globe?

20 years... I'm safe.
 
Henning,
your comments about few decision needing to be made and how humans fail,at them more than not .......are honestly ridiculous and disrespectful to professional aviators...87,000 flights take off and land each day in this country.....being on an airplane is just about the safest way you can spend your time statistically....
 
Last edited:
44ae2003_bth_threadderailedsu2.gif
 
Henning, it is you who are in denial. Did you even read my post #614? Any answers as to the reality of implementing this system? To the cost and logistics involved? Could you get an autonomous aircraft to takeoff from Point A and fly to Point B in benign conditions, sure. But to replicate that for 100,000 flight each day across the globe?

20 years... I'm safe.

Ok...
 
Here is another example of where it's going human wrong. Here's another Italin captain ready to pull a Costa Concordia.

http://youtu.be/RWetojC0ul0

Here's a computer screwing up a rocket...

I know it's not a direct comparison but it can happen. They're only as good as their sensors and the programmers who wrote the code.
 
Henning,
your comments about few decision needing to be made and how humans fail,at them more than not are honestly ridiculous and disrespectful to professional aviators...87,000 flights take off and land each day in this country.....being on an airplane is not the safest ways you can spend your time statistically....

The decisions are all simple logic decisions with limited and with databasable field of options that is larger than the human mind can handle, yet a computer can refrence through in seconds, and with much greater accuracy than the human mind is capable of.

Cost of implementing is not an impediment, it will be a profitable investment at nearly any cost.

I am a pilot, and I am an airline service consumer, and I would prefer an autonomous airliner at this point. That the general public favors my opinion more than yours is something I would wager a significant amount on.
 
And due to the desire to cut costs.

BTW, the company leading the way in AI is Google - it's no secret that they seek to develop and implement AI. One of the founders fathers was involved in AI research.

Think of that every time you let them place a cookie, scan your email, or track your android phone everywhere you go.

Another reason to hate Google. :mad:

Cookie Culler for Pale Moon (and presumably Firefox) has a nice feature that will delete all "unprotected" cookies every time you open the browser.

Rich

(Posted from my BlackBerry Q10)
 
I love computers, I love automation... I used to be a programmer. Computers are great at doing simple predictable tasks with great precision, speed, and lack of errors. However, they're still really really bad at decision making. I will grant you it is getting pretty good these days but they have no ability to go beyond scenarios/parameters laid out by the engineers and programmers who designed them.

I don't even mind the idea of the computer driving or flying, they are less likely to make a mistake. However, I still want a human in the loop somewhere in case the machine does something stupid or creative solutions are called for.
Driving a car is hard for a computer. Computers flying a plane is easy. Stupid human tricks have wrecked enough airplanes it is time to give computers the chance to prove their mettle(go ahead post the airbus runway flyby crash.:rolleyes:)
 
Henning, it is you who are in denial. Did you even read my post #614? Any answers as to the reality of implementing this system? To the cost and logistics involved? Could you get an autonomous aircraft to takeoff from Point A and fly to Point B in benign conditions, sure. But to replicate that for 100,000 flight each day across the globe?

20 years... I'm safe.

You're missing the point.

Henning is not a tech expert. Far from it. He's a user with no experience in the field, who sometimes confuses the field with science fiction.

Those of us who actually do develop control systems aren't very quick to jump on the AI airliner bandwagon. Heck, I stare at the "almost there" Google autonomous cars every day out my office window. And I have for years now. It's no accident that they all have two people in them…

When you have working autonomous cars and then busses, you might consider thinking about airliners. But working autonomous cars are several years away if they ever come -- and I really don't think they will.

What is substantially more likely is autonomous assist, but we already have that.

As was said earlier, the issue is the failure modes. If you can't enumerate them somehow, you can't account for them.

I find it really hard to believe that an AI system could have averted the Air France problem. That, as for many other airliner accidents, was due to a failure in the automation. You don't solve that with more automation. You make it worse.
 
The decisions are all simple logic decisions with limited and with databasable field of options that is larger than the human mind can handle, yet a computer can refrence through in seconds, and with much greater accuracy than the human mind is capable of.

Where does that "database" come from?

Observed faults? Guaranteed incomplete. People will die.

Analysis? You had better hope the analyst thought of everything. I've been through enough design cycles to know that ain't happening.
 
Here's a computer screwing up a rocket...

I know it's not a direct comparison but it can happen. They're only as good as their sensors and the programmers who wrote the code.

Of course computers can screw up, but humans screw up at a much higher rate now. As soon as the insurance industry decides that actuarially pilots are the higher risk factor, they will be replaced. Guess what? They figured that out a couple of decades ago, the system of implementation has been in progress since, and we are closing in on that time.

This will have nothing to do with egos or abilities, it is strictly a financial decission by the people who have Strict Liability exposure. All you have to do is follow the money. Computers cost both the insurance industry and the operators less money than pilots, ergo pilots will be replaced by mechanization. First it was the slaves, next it was the factory worker, next it was the company switch board operators, secretarial pools, and Mail rooms. Office worker productivity increased and staffs reduced. Travelling salesmen, warehouse workers and fork lift operators. Even brain surgery is now done robotically. If you think being a bloody airline pilot is more difficult than being a brain surgeon, you're delusional.
When people get surgery now, they look for the robotic machines, and all the best surgeons are buying them. Take the hint.
 
Driving a car is hard for a computer. Computers flying a plane is easy. Stupid human tricks have wrecked enough airplanes it is time to give computers the chance to prove their mettle(go ahead post the airbus runway flyby crash.:rolleyes:)

As I said, I think they have their place and under normal conditions just playing the odds it's probably safer.

I'm just not going to ride in something- airplane or car that doesn't have an override button and a human who can take control. I've worked with too many electronics/computer systems and seen them do too many crazy things under different failure scenarios.
 
You're missing the point.

Henning is not a tech expert. Far from it. He's a user with no experience in the field, who sometimes confuses the field with science fiction.
Yeah, I'm getting that...

I realize that he's an expert on lots of things, whether it be talking down a suicidal caller late at night from his childhood bedroom phone, or the prayer language of Islamic suicide bombers.

I know close to nothing on lots of things, but one thing I know a little about is flying large multi-engine aircraft around the globe, and everything that's involved with that. And I'm pretty sure I have Henning beat there.

I agree with you (and you seem to be the expert here) that autonomous airliners are science fiction for now, and probably will be for a long time.

You're also right, that we're closer to assisted flying, but with all the redundancy you're going to have to build in for single-pilot ops, you might as well keep the second guy (or gal).
 
Back
Top