Moderators: jsumali2, richierich, ua900, PanAm_DC10, hOMSaR

 
N312RC
Posts: 2615
Joined: Mon Aug 28, 2000 10:58 am

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 5:50 am

anshabhi wrote:
2 people for a machine carrying 200+ lives is too expensive for airline industry huh?

Even today, Airlines barely get any sympathies from the general public due to charging for petty things. This will cause a complete mistrust between people and airlines.


Then the general public of which you are are part needs to stop asking for $3.50 JFK-LHR fares? Airline employees have to live too... must be nice to be able to live in an alternate reality.

Then again this is airliners.net, the place where people apparently love the airlines, just hate the people that work there/run them/the cost of flying/the fact that no airline is flying a 77W to DSM/etc etc. Eye roll.
 
BA777FO
Posts: 577
Joined: Sun Apr 22, 2018 2:58 pm

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 5:55 am

kalvado wrote:
BA777FO wrote:
kalvado wrote:
And are you sure you decision on KEF vs YYT is the best one? (nothing personal, "you" = human pilot). If you made it to the runway, it was probably a good one anyway. The difference in terms of risk is probably not that major if you can fly another 2 hours to diversion point. Yes, regulations require minimal flight time and all that, but how big it actually is? You wouldn't hesitate between 1 and 3 hours flight time, it may be 95 vs 90 minutes where you would think twice. As for the flight path taking wind into account - computer may have a wind map, and I would certainly do a fine job calculating the times.

As for the best option... Did you ever had to look back at your own actions and say "well, I would be much better off doing things differently.." ? I certainly did, both personally and professionally.

Moral of the story - there are situations where human brain is better than machine, but those become fewer as time goes by. I wouldn't look for examples of human superiority in situations covered by checklists. It is unknown unknowns when humans are better. And it turns out that these are the situations where human can screw up.


This isn't about human superiority vs. checklist use by a computer - it's about a situation that a computer cannot diagnose because of its multi-faceted complexity.

As for the KEF/YYT situation that's why we use a decision making tool - TDODAR. How well can a computer generate options, make a decision and, as you refer to, review your decision. That's what the R is - review, we do it after taking the decision and its a continuous process all the way to a safe outcome. How well can a computer change its decision based upon new, previously unknown information?

Maybe we end up with the remote pilot drone system decades down the line but then you have a bunch of issues around interference etc.

You're talking about multiple-choice decisions based on a closed set of information (divert- YES/NO; if YES - which of 3 nearby airports best fits the scope- based on distance, weather, availability of medical facility, ARFF) and effectively an algorithm for taking such decision. This is type of task where a computer can - and eventually will - excel humans. Feeding new information, if any, in the same algorithm is basically the review portion.

Basically I would say if you can envision something being a problem - it can be programmed along with resolution measures. It is situations of "who could have thought of it??" type that would matter.


It's not necessarily closed information, though. It's all well and good another poster saying there'll be even more sensors and resundancy in the unreliable airspeed scenario but most of those raised would be useless anyway. Let's say the technology is found, how does the computer know which reading, if any, to trust? When it trusts none of them, how will it proceed? The one thing with flying is that things are constantly changing. Based upon known information an algorithm says divert due to a technical fault. However, coupled with a fuel leak and weather avoidance half way along it determines it'll be at fuel exhaustion by the time it arrives, what does your algorithm say then? The computer will make decisions based upon what it knows now - it knows it has a current fuel state and can predict what its burn will be to destination. Can it factor in an unknown rate of loss into its decision when it can't determine whether it has a genuine leak or a faulty guage?

Maybe it can do all of that. I doubt it'll be able to do that in my kids' lifetime with any degree or reliabity let alone my own.
 
FluidFlow
Posts: 710
Joined: Wed Apr 10, 2019 6:39 am

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 5:56 am

BA777FO wrote:
FluidFlow wrote:

Look I respect what you are doing but all you do is defend your job on a 777, an archaic aircraft that needs a pilot to function.


I'm not defending my position - there are just some subjective situations that computers cannot determine the most appropriate course of action due to cascading failures or are unable to diagnose/recognise the problem. You mention GPS for speed - that'd be groundspeed, not airspeed, quite a key difference.

You talk about a bunch of sensors that can all be used - at the moment we can fly with deferred defects - the computerised aircraft won't be able to. That's a lot of extra maintenance and cancelled flights. The example you used about shutting down airspace around volcanoes will shut aviation down for weeks when we're now able to operate safely much closer to active areas.

It's nice that you think all of these issues can be overcome, maybe it will, but like I said, today's most advanced technology is just being delivered with a 30 year lifespan and the most advanced that has moved on is auto TCAS. You have a long way to go before it can make complex decisions.


And i agree with you it will not be here in the next 15-20 years and even in 40 years a lot of aircraft will still need pilots but the transition will start really soon.

As you say GPS is ground speed, but if you have GPS + the actual weather and wind information from the local area based off all the other aircraft and surveillance vessels (satellite, weather balloons, etc) you get a really accurate state of the atmosphere and the computer can then determine what is right and what is wrong.

One thing that will come first and I am pretty confident that it will be in the next clean sheet design is that after every flight all the information from the flight recorders will be read out and used to train AI-pilots. If you do this over years you get millions of flights and cover 99% of incidents that occur rather frequently. This is the base to establish full auto pilots used for single pilot operation. Then from there you go to self flying aircraft. That is 40-60 years down the road but if I am lucky I might be able to take a flight on one of those. I would be around 100 years old but I hope I make it there.
 
Speedalive
Posts: 166
Joined: Sun Jan 17, 2016 7:09 pm

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 6:05 am

NASA Chief: Uncrewed Aircraft 'Safer'

https://www.ainonline.com/aviation-news/general-aviation/2020-07-27/nasa-chief-uncrewed-aircraft-safer?utm_hsid=45278629&utm_campaign=AIN%20Alerts&utm_medium=email&_hsmi=92182624&_hsenc=p2ANqtz-81JXGiBqJn_VokEEBpNIbCpuS-DekciTaXuBWU02h6XtoW331wKDAuYQKufAeb3qWNqwnWmi6dPsYi862PKsjN9l9_8Q&utm_content=2&utm_source=hs_email

As a commercial pilot just getting started in the industry, I wonder how I could get ahead of this and thrive in a situation where my job is replaced by computers. It’s not worth arguing against it - this amazing job will disappear in due time. Think about the progress in aviation over a hundred years. It’s really not far fetched to think that we can be replaced by advanced AI in a much smaller time frame than many people on here are saying. Especially when there’s a profit incentive. Cut the pilot labour costs, increase aircraft pax capacity, reduce flight delays during IROPS and associated costs, maintenance, etc.. Airline management will be all over this.

What kind of education would be beneficial to have to take advantage of this trend towards increasing automation? Learn to become a programmer? Become an aerospace engineer?
 
User avatar
Airbuslightie
Posts: 4
Joined: Wed Jul 29, 2020 1:13 pm

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 6:10 am

MillwallSean wrote:
Am I missing something here? we are arguing that no technological progress should happen and that two person flying aircrafts should, forever, be the norm. Hmm thats not how the world works. 30 years ago there were 3 persons flying airplanes. 50 years ago there were a team.
10 years from now it will be one person and the backup will sit in a boring suburban office, stepping in should something happen to the pilot (see dead-hand system on bullet trains etc).
I can guarantee that the flying public wont care whether there is 1 or 2 persons flying the plane just as they didn't care when we went from 3 to 2 persons. They will however care when we take the step to fully automated planes without pilots. That is a huge step, most of us humans wants safety / redundancy and we want to know that there is a human that can step in and think, not just analyse. But we don't need two humans to think, one is perfectly fine. And cockpits can be designed for one person these days, the tools, the help is more than satisfactory.

For me this step isn't so much a technological one, its more an institutional one. Unions etc will fight tooth and nail against it, trying their best to drum up support for maintaining status quo. Its a lost battle but I am sure that it will delay the implementation in certain jurisdictions. But for me, bring it on as humans we need to continuously evolve and improve and one person cockpits is a given on at least shorthauls.


Some very good points.....especially about technological progress. The argument here is less about the progress as I have seen some nice progress on the A380. We now have auto TCAS and BTV, both brilliant and make my job easier. Much of the tech suggested here required for autonomous flight will need to pass a real world test. and incrementally. That means that I, as a pilot, need to see and use this technology. Since I am not seeing much of it, that implies possibly a decade before proper testing even begins. I do look forward to my job becoming incrementally easier but to date I am presented with massive challenges every time I enter the simulator, and these challenges stem from issues that the aeroplane cannot resolve without human intervention.

I would also disagree with your assertion that we will have single pilot aeroplanes in 10 years. In order to have single pilot aeroplanes, we need to evolve an additional iteration of technology. In other words, that plane must be able to operate with no pilots. I take your point about the pilot on the ground(one would have to ask what is the point) but the comms would have to be bulletproof. Given the fact that Garmin can't even keep their website up and running, clouds that prospect quite significantly.

What we might see in 10 years is 2 pilot ultra long haul flights, instead of the current 4 pilot compliment.

You are probably right about some pushback being about protecting jobs and there is nothing wrong with that. It serves as a useful "peer-review" for the concept. If the concept is solid, it should pass any objections from unions and the like. Just because someone has an agenda, does not make them wrong.
 
willfinn
Posts: 4
Joined: Fri Mar 20, 2020 10:33 am

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 8:39 am

This has been one of the more interesting threads. From the discussion, it seems that in order to have commercial airliner UAVs, we'd need at least the following:

Complete redesign and reworking of ATC worldwide

Completely new approach to commercial aircraft architecture on all levels of design and manufacturing

Significant liability assumed by OEMs. Pilots would no longer assume primary responsibility for safety. After all, they are paid, because they take responsibility? Safety would be in the hands of software engineers and programmers. (Incidentally, the hull loss accident rate of military UAVs is high.) If a product malfunctions, the manufacturer are responsible. Who is responsible for (eventual) loss of life?

Reliable, uninterruptible communication between ground control and the aircraft

All of the above would have to be extensively simulated, tried and then implemented. It won't be enough to punch in a few lines of code and strap some extra sensors on existing aircraft.

This would also entail a complex, interacting, network of AI driven systems. This might very well be so complex, and so difficult to maintain, that in order to achieve efficiency and redundancy it would require more human intervention than the "dumb" systems of today.

A 10-year time frame is unrealistic. 50 years is such a time frame that there will be no guarantee of anything, let alone ROI.

I fully understand, and support, the arguments brought forth by several knowledgeable posters, but I remain doubtful: Who is going to pay for all this? The OEMs are struggling, not to mention the airlines.
 
whiplash
Posts: 76
Joined: Mon Nov 14, 2011 12:47 am

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 9:55 am

Here we go again

In aviation, everything is about money and how to save every last penny! Having been through most of the replies on this forum, it's laughable to see that people simply don't understand the concept of what can and can not be installed in an airplane. The simple number of cameras, sensors, redundancies, computers, and all the other thingamajigs that will have to be installed in each airplane to be able to decipher all the 10,000 things that can go wrong with it will add to the cost. An extra $10 million an airplane for this technology will not be the cost-saving that the airlines will be looking at. Let's agree that airlines are not paying $10 million in salaries over a period of 30 years (lifetime of the aircraft).

Let's for a second assume the price is not an issue. The ridiculous amount of redundancies that we would have to install in these planes to make them completely autonomous will also render these planes susceptible to AOG thanks to something called the MEL. As the automation increases, the minimum equipment list of the aircraft will grow as well. So for a fact, without the #5 right fuselage camera, the aircraft will be unable to taxi out and will need to be replaced before the aircraft can be dispatched. Even worse if the #5 camera dies in flight, making sure the aircraft will simply refuse to vacate being blind on one side. That would be a hell of a day for the controllers. No matter what anyone says, it's fairly common for things to fail on a daily basis in an airplane. Nobody would want a plane that gets grounded for the tiniest of fault in its systems.

Fatigue management is a joke as it is right now when it comes to commercial flying. Duty hours for pilots already cross 12 hours easy with long flying times. With a single pilot operation, these duty hours and flight times will have to be reduced significantly (especially in short-haul operations). Considering that I do a 4 sector flight on a daily basis, I'd be tapping out at 2 sectors with single-pilot operations. The other guy would be filling my seat to do the remaining 2, which would make very limited difference when it comes to the pay.

You will also not be getting a lot of real estate if the aircraft becomes autonomous. The cockpit or some form of it will continue to be present in the airplane for at least a rudimentary backup and/or troubleshooting.
A320 driver
 
Galore
Posts: 27
Joined: Fri Apr 17, 2020 12:43 pm

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 12:07 pm

Gremlinzzzz wrote:
anshabhi wrote:
If every industry was not a race to the bottom, then airlines would not need to charge for petty things.If unions were not always pushing for higher pay, or management taking money out instead of paying down debt.........you get the drift, then maybe we would get more healthy airlines.


Not every industry is a race to the bottom. Consumer tech is always getting better and cheaper. Not worse and cheaper (compare a TV or a cell phone from now to 1985). Restaurants (at least where I live) are not getting worse, cars are getting better, grocery stores have a lot more selection and higher quality than 20 years ago and so on.

The fairly unique quality and supply problem with airlines is that coach class races to the absolute tolerable bare minimum to squeeze out the last cent while something like “main cabin extra” costs a totally unreasonable extra for the same seat in a “better” location. The only airline that doesn’t suffer from this lack of quality options for a justifiable up charge ironically is Spirit with their BFS product.

As long as they themselves offer a massively (!) cheaper seat in the same plane then it’s frankly financially irresponsible to pay extra for one a few rows up with 3 inches more legroom (or 3x for a first class seat). The value proposition of the “better” seats is simply not there (other than BFS). If AA would offer BFS for $50 extra like Spirit (without upgrading their thousands of frequent fliers for free), they’d sell out all the time.
 
TObound
Posts: 781
Joined: Mon May 27, 2019 12:54 am

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 12:30 pm

The comments here are naive. The amount of people who think automation can't replace aircrew sounds like every other sector and job before they were automated out.

Sure it may not happen soon. And it will be a process. But it is coming. The speed with which AI can analyze and make decisions, along with the cost of pilots, and the relatively lower workload at cruise (ie. Lower productivity) is exactly why their jobs are targets. Even eliminating a relief pilot or one in the cockpit is a huge savings.

Want to know who has already been through the process? Military combat aircraft. Look up how checklists work in an emergency on the F-22. And those have to be designed for combat damage. Automation and improved sensors has enabled single seat cockpits for air forces and made those aircraft more capable over the years. There is no reason the same can't be done in the transport sector.
 
kalvado
Posts: 2863
Joined: Wed Mar 01, 2006 4:29 am

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 12:54 pm

BA777FO wrote:
kalvado wrote:
BA777FO wrote:

This isn't about human superiority vs. checklist use by a computer - it's about a situation that a computer cannot diagnose because of its multi-faceted complexity.

As for the KEF/YYT situation that's why we use a decision making tool - TDODAR. How well can a computer generate options, make a decision and, as you refer to, review your decision. That's what the R is - review, we do it after taking the decision and its a continuous process all the way to a safe outcome. How well can a computer change its decision based upon new, previously unknown information?

Maybe we end up with the remote pilot drone system decades down the line but then you have a bunch of issues around interference etc.

You're talking about multiple-choice decisions based on a closed set of information (divert- YES/NO; if YES - which of 3 nearby airports best fits the scope- based on distance, weather, availability of medical facility, ARFF) and effectively an algorithm for taking such decision. This is type of task where a computer can - and eventually will - excel humans. Feeding new information, if any, in the same algorithm is basically the review portion.

Basically I would say if you can envision something being a problem - it can be programmed along with resolution measures. It is situations of "who could have thought of it??" type that would matter.


It's not necessarily closed information, though. It's all well and good another poster saying there'll be even more sensors and resundancy in the unreliable airspeed scenario but most of those raised would be useless anyway. Let's say the technology is found, how does the computer know which reading, if any, to trust? When it trusts none of them, how will it proceed? The one thing with flying is that things are constantly changing. Based upon known information an algorithm says divert due to a technical fault. However, coupled with a fuel leak and weather avoidance half way along it determines it'll be at fuel exhaustion by the time it arrives, what does your algorithm say then? The computer will make decisions based upon what it knows now - it knows it has a current fuel state and can predict what its burn will be to destination. Can it factor in an unknown rate of loss into its decision when it can't determine whether it has a genuine leak or a faulty guage?

Maybe it can do all of that. I doubt it'll be able to do that in my kids' lifetime with any degree or reliabity let alone my own.

Well, you're talking as if this is the first time people are dealing with complex systems.. On of the approaches is to make computer to be "aware" of situation - probably more so than a human pilot.
For example, there can be a predictive model running taking all possible data as inputs and predicting what would happen in the next minute. It can take into account a lot of things and compare a lot of things. Something unexpected happens, like airpeed on one sensor dropped - but that happened? Did engine thrust go down? Did AoA change? Is climb/descent speed consistent with airspeed and AoA? What about GPS speed and altitude - are they changing in tune? Are things diverging(most likely hardware issue) or converging to a new meaningful situation (maybe windshear?) Which input(s) doesn't fit the picture?

This is similar to the way human pilot would think about it - taking the bigger picture. And yes, a computer can do it better since it can "remember" all the parameters for the past minute in fine details, including calculated wind speed, small deflections of controls and what not. If you will, "unrealiable airspeed" procedure this setting pitch and thrust is on the same page - going to a predictable state and checking out the outcome. Except for computer can have many more "memory item" states - actually all of them.

True, putting such a system together is a very non-trivial task. But we're talking about multi-billion companies investing tens, hundreds, maybe thousands of man-years into solution. They can do much better than someone would in 3 minute forum post
 
User avatar
Airbuslightie
Posts: 4
Joined: Wed Jul 29, 2020 1:13 pm

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 2:07 pm

I think, perhaps, we are all going off on a tangent here.....

The last line of the article states the following :-

For its passenger jets, though, Airbus states the tech won't replace pilots in the cockpit but will make flying safer by helping reduce workload.


It would appear that even Airbus appreciates the challenges of removing pilots from the cockpit and that is not the goal. The idea is to increase automation and reduce workload. Presumably the pilot will fulfill the redundancy roles, a vastly different proposition to removing all human presence in the flightdeck.
 
kalvado
Posts: 2863
Joined: Wed Mar 01, 2006 4:29 am

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 2:27 pm

Airbuslightie wrote:
I think, perhaps, we are all going off on a tangent here.....

The last line of the article states the following :-

For its passenger jets, though, Airbus states the tech won't replace pilots in the cockpit but will make flying safer by helping reduce workload.


It would appear that even Airbus appreciates the challenges of removing pilots from the cockpit and that is not the goal. The idea is to increase automation and reduce workload. Presumably the pilot will fulfill the redundancy roles, a vastly different proposition to removing all human presence in the flightdeck.

Single plot is a much more feasible goal, and you can honestly say that means "not removing pilots". Besides, it doesn't have to be a quantum leap "fire everyone tomorrow!". Think about it as a next generation autopilot with vastly greater functionality. Once technology is deployed and flown for millions hours, there will be much more substance in discussion.
 
BA777FO
Posts: 577
Joined: Sun Apr 22, 2018 2:58 pm

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 2:31 pm

Given that they've just figured out auto-following of a TCAS RA, I think what you describe is light years away. Like I said initially, it won't be in my lifetime, probably not even my kids' lifetimes either.
 
User avatar
CarlosSi
Posts: 657
Joined: Sat Jul 08, 2017 8:29 pm

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 2:40 pm

IMO there should always be a backup pilot in the cockpit for the same reason we will likely always have two engines as the fewest amount of engines on part 121.
 
travelsonic
Posts: 106
Joined: Sat Apr 07, 2007 8:59 am

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 3:08 pm

IADFCO wrote:
One can envision potential solutions to all three, but in any case the computational power required is way beyond what is practical today. .


Is it that far off though, given how smaller and smaller increasingly powerful computers are getting?

I mean, I guess one part of it (of many parts) would be how much of it can be parallelized - that is, how many tasks can be broken up and distributed across multiple cores/threads, how much data can be spread across that load to speed up the processing of all that information.
Given just how many cores/threads that processors (especially server and workstation oriented CPUs) can have these days, the more that could be parallelized, the better in terms of getting through, and processing lots of data / variables at once.

Personally, I think that (at least for some time after this tech becomes more ready for applied use) still having people in the cockpit is a must.
 
kalvado
Posts: 2863
Joined: Wed Mar 01, 2006 4:29 am

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 3:24 pm

BA777FO wrote:
Given that they've just figured out auto-following of a TCAS RA, I think what you describe is light years away. Like I said initially, it won't be in my lifetime, probably not even my kids' lifetimes either.

My personal experience - whenever I say "I will NEVER do that!" - I will be doing that exact thing full time in 5 years.... Happened more than once to me.
 
BA777FO
Posts: 577
Joined: Sun Apr 22, 2018 2:58 pm

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 3:28 pm

kalvado wrote:
BA777FO wrote:
Given that they've just figured out auto-following of a TCAS RA, I think what you describe is light years away. Like I said initially, it won't be in my lifetime, probably not even my kids' lifetimes either.

My personal experience - whenever I say "I will NEVER do that!" - I will be doing that exact thing full time in 5 years.... Happened more than once to me.


Given that they're just delivering what is now state of the art A350s and B787s with 30 year lifespans nowhere near capable of the nirvina you envisage I reckon I'll gamble I'll be doing what I'm doing when I retire in 30 years time!
 
kalvado
Posts: 2863
Joined: Wed Mar 01, 2006 4:29 am

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 5:44 pm

BA777FO wrote:
kalvado wrote:
BA777FO wrote:
Given that they've just figured out auto-following of a TCAS RA, I think what you describe is light years away. Like I said initially, it won't be in my lifetime, probably not even my kids' lifetimes either.

My personal experience - whenever I say "I will NEVER do that!" - I will be doing that exact thing full time in 5 years.... Happened more than once to me.


Given that they're just delivering what is now state of the art A350s and B787s with 30 year lifespans nowhere near capable of the nirvina you envisage I reckon I'll gamble I'll be doing what I'm doing when I retire in 30 years time!

I wouldn't say "nowhere near" actually. I may be wrong, but since those are FBW planes, they shpuld have most connectivity in place. I can envision significant radio upgrade when - and if - ATC would get out of analog transmission stone age applied to many existing planes, and that would be another big missing piece. Conceptually, autopilot already has a lot of control over the plane, and replacing it shouldn't require serious redesign.
Of course, going switch by switch there will be a lot missing - for example I assume landing gear, air conditioning system, fire bottles, engine start/stop are all manual only...
And since my personal expectation is a single pilot operation in foreseeable future, packs may stay manual to keep that sole pilot busy. If I was a betting man, I would offer you $20 bet on you flying as a sole pilot / captain before you retire.
UPD: to put my 30 year expectations in perspective - I believe A320 was the first FBW passenger aircraft, and the model is less than 35 year old...
 
bhill
Posts: 1837
Joined: Thu Sep 13, 2001 8:28 am

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 6:17 pm

Right up until the next class X solar max storm..or EMP, or insurance rates going through the roof. Although,I have wondered what any full FBW aircraft is able to handle EMI now...
Carpe Pices
 
User avatar
flybynight
Posts: 1535
Joined: Wed Jul 30, 2003 1:58 pm

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 6:19 pm

Waterbomber2 wrote:
If they can make cars drive autonomously, why wouldn't they be able to make aircraft fly autonomously? Less obstacles, less improvisation required.

Pilots still have to save the day almost everyday with current tech aircraft but checklists can be run by computers too and faster.
It all is a matter of how many scenarios you can program into the computer, so what OEM need to do is to collect data from as many flights as possible, and program all the scenario's so that issues can be resolved autonomously.

An onboard pilot serves as back-up, but the airplane should also be able to be flown remotely.

Pilots hate to talk about this, but the fact is that it only takes one mistake, one bad judgement, and your children and your passengers' children may never see their mommy or daddy ever again.
This is now a rare occurrence, but if we can do even better, why not?
We are only few years away from launching to Mars with hyper-automated machines, so there'll be jobs for pilots, and they'll be more meaningful and satisfying, no need to be protective.
Let's embrace progress and focus on bigger things.


Well first of all, autonomous cars aren't doing as well as originally thought. A lot factors have come up to put a damper on it.

Humans can predict based on data that computer can never factor in. Could a computer "think" that child on the side of road looks like he/she might jump into the road based on experience and the way the child is behaving?
Also if we go to one pilot, what if he/she becomes sick?
The less pilots fly the plane, the less experience they also rack up. I want my pilot sharp and ready to handle situations. Not as a backup to IBM.
Heia Norge!
 
airzona11
Posts: 1756
Joined: Wed Dec 17, 2014 5:44 am

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 7:19 pm

flybynight wrote:
Waterbomber2 wrote:
If they can make cars drive autonomously, why wouldn't they be able to make aircraft fly autonomously? Less obstacles, less improvisation required.

Pilots still have to save the day almost everyday with current tech aircraft but checklists can be run by computers too and faster.
It all is a matter of how many scenarios you can program into the computer, so what OEM need to do is to collect data from as many flights as possible, and program all the scenario's so that issues can be resolved autonomously.

An onboard pilot serves as back-up, but the airplane should also be able to be flown remotely.

Pilots hate to talk about this, but the fact is that it only takes one mistake, one bad judgement, and your children and your passengers' children may never see their mommy or daddy ever again.
This is now a rare occurrence, but if we can do even better, why not?
We are only few years away from launching to Mars with hyper-automated machines, so there'll be jobs for pilots, and they'll be more meaningful and satisfying, no need to be protective.
Let's embrace progress and focus on bigger things.


Well first of all, autonomous cars aren't doing as well as originally thought. A lot factors have come up to put a damper on it.

Humans can predict based on data that computer can never factor in. Could a computer "think" that child on the side of road looks like he/she might jump into the road based on experience and the way the child is behaving?
Also if we go to one pilot, what if he/she becomes sick?
The less pilots fly the plane, the less experience they also rack up. I want my pilot sharp and ready to handle situations. Not as a backup to IBM.


With connectivity, an incapacitated pilot can mean remote take over and descent/land at closest alternative airport. That is within the realm of today's technology. As for the analogy of predicting what the child will do... a computer can make assessments on movement on all sides of the car, from farther distances away than humans. Sure if the driver is solely staring at and aware of that one child, they can react, but automatic breaking for example is a technology that has successfully rolled out to most new model cars to prevent what humans might miss or slowly react to. Insert adding more automation into the cockpit and making flying safer.
Last edited by airzona11 on Thu Jul 30, 2020 7:23 pm, edited 1 time in total.
 
ukoverlander
Posts: 406
Joined: Thu May 13, 2010 10:57 pm

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 7:23 pm

cirrusdragoon wrote:
https://www.businessinsider.com/airbus-completes-autonomous-taxi-take-off-and-landing-tests-2020-7

It will be interesting to see how this develops over ten , twenty years from now . Will there be a reduction in pilots onboard? Could society be persuaded to welcome a pilotless flight deck? Time will tell.


In 20 years time Boeing will roll out the pilotless '737 NG MAX NEO Next Generation Warm Over' with the very latest evolution in overhead bin, cabin lighting, cushionless seat and tiny cramped cockpit technology.
 
peterinlisbon
Posts: 1811
Joined: Wed Feb 01, 2006 3:37 am

Re: Airbus completes testing of self flying jet

Thu Jul 30, 2020 8:37 pm

They should add a function that says "hey melon, what about the wheels?" and then takes over if you try to land on the engines at 200kt.
 
Gremlinzzzz
Posts: 90
Joined: Fri Jan 24, 2020 4:28 am

Re: Airbus completes testing of self flying jet

Fri Jul 31, 2020 7:40 am

Galore wrote:
Not every industry is a race to the bottom. Consumer tech is always getting better and cheaper. Not worse and cheaper (compare a TV or a cell phone from now to 1985). Restaurants (at least where I live) are not getting worse, cars are getting better, grocery stores have a lot more selection and higher quality than 20 years ago and so on.

The fairly unique quality and supply problem with airlines is that coach class races to the absolute tolerable bare minimum to squeeze out the last cent while something like “main cabin extra” costs a totally unreasonable extra for the same seat in a “better” location. The only airline that doesn’t suffer from this lack of quality options for a justifiable up charge ironically is Spirit with their BFS product.

As long as they themselves offer a massively (!) cheaper seat in the same plane then it’s frankly financially irresponsible to pay extra for one a few rows up with 3 inches more legroom (or 3x for a first class seat). The value proposition of the “better” seats is simply not there (other than BFS). If AA would offer BFS for $50 extra like Spirit (without upgrading their thousands of frequent fliers for free), they’d sell out all the time.
A race to the bottom is essentially a situation where competing companies are trying to offer a service at ever decreasing prices. Aviation is the very definition of this, and airlines are at times doing this not to their benefit.

If an airline tries to keep the same product it had in the 70's i.e. seat pitch, service, get rid of seat selection fees, keep service as it were, it would not be competitive. They reduce the price of the ticket and look to make revenue elsewhere because this is what the market demands.

In aviation, you needed to go to a travel agent, you needed someone to check you in, a plane used to have a navigator, a flight engineer. Today, you can book a ticket by yourself and bypass the travel agent, something that has brought prices down. Low cost airlines have got rid of a lot of employees that used to traditionally work for airlines at airports. Today, with things like digital twins regarding engines, you do can diagnose most issues without needing to pull out an entire engine. Aviation tools and automation are getting better, and eventually, we will get to a position where airlines, which have struggled to rein in crew costs, especially pilots because of upward market pressures, replace them the moment they can.

So, when they eventually have an option to automate, they do that.
 
Sokes
Posts: 1678
Joined: Sat Mar 09, 2019 4:48 pm

Re: Airbus completes testing of self flying jet

Sun Aug 02, 2020 3:41 am

IIRC if one writes vocabulary twenty times, new neuronal structures build up. It's not only uploading a file, it's hardware change.

I don't remember the exact number, but Elon Musk said something like it's easy to get the computer make only ten mistakes in one billion operations, but difficult to get it make only one mistake.
But then are humans perfect?

Russia had a space based missile warning system. One night the system showed an attack by a single missile. However why start a nuclear war with a single missile?
The man in charge decided it has to be a malfunction, ignored orders and didn't disturb his superiors sleep. It turned out a reflection of sunlight caused the malfunction.

A single pilot with authority to override the computer is like a prime minister with authority to override parliament.

One has to start with cargo. Once computer failures are less likely than failure by two pilots one can do it.
Why can't the world be a little bit more autistic?
 
strigon07
Posts: 2
Joined: Tue Mar 17, 2020 8:56 am

Re: Airbus completes testing of self flying jet

Sun Aug 02, 2020 10:25 am

Been lurking on this forum for a while now, finally thought I’d chip in. To start, I have no doubt that increasing automation, if done correctly, can help to significantly improve the safety in the industry. And I do believe that there is an economic gain to be had too (beyond the examples already shared) - in certain phases of flight, e.g. cruise, take-off, and landing, automation may allow for higher efficiency by selecting the optimum or near-optimum configuration for the required task. Computers are good at handling a large amount of information (dealing with lots of parameters) at any point in time, so monitoring of complex systems is very much up their alley, particularly when in the "normal" operating window.

However, I have my concerns about the challenges of full automation (i.e. completely pilotless aircraft). Let's start with the technical problems.

Firstly, the need for communication. As some have already pointed out, a complete digitalization of existing ATC infrastructure is required for fully autonomous aircraft. There are significant regulatory, economic, and technical challenges associated with this. Since such a system would be able to direct the direction of an aircraft, it must be secure against potential nefarious actors.

In the case of emergencies (perhaps medical), such a communication system needs to be flexible enough for on-board crew to communicate with ATC, e.g. to seek advice from/inform emergency services. Again, such a contextually-aware communication system has to be robustly implemented and tested for emergency use, which will be… non-trivial. A fallback to a voice-recognition based system is possible, but also problematic. For one, there’s the issue of reliability and overall contextual awareness. Can this be done? Maybe – but there are computational limits and difficulties associated with the type of deep-learning networks that are often employed for these types of tasks.

There is also the issue of dealing with “unknown-unknowns” in the case of emergencies. This has traditionally been an area where AIs have been not so great at. Autonomous systems may be unable to improvise during emergencies, which may lower the chances of survival. For example, consider the Air Astana incident with cross-rigged controls. The pilots were about to regain some control by learning, quite literally, on-the-fly. Many machine learning systems often require hours of supervised learning in order to generate a model capable of responding in an acceptable manner, so dealing with reduced/altered functionality in a dynamic environment would again be difficult. (Of course, for argument’s sake – yes, an autonomous system might have been able to detect the maintenance fault before the pilots ever did, hence my preamble. But this example serves to highlight a specific challenge that engineers/computer scientists will have to grapple with – the challenge of developing a system that can deal with situations that are radically different from what might be expected.)

Then comes the problem of trust, especially when dealing with complex deep-learning algorithms. From a pilot’s perspective, when making certain decisions, pilots adopt a rule-based framework. Such a framework is rooted in a logical understanding of cause and effect, and on an intuitive understanding of the principles of flight and the physical processes that power all of an aircraft’s systems. Deep-learning models, however, do not always rely on such rule-based frameworks. Instead, they rely on brute-force association with a lot of sample data. So there is often an issue of generalizability and reliability in fringe scenarios. Quite a number of my university’s physics profs have questioned the “validity” of using deep-learning models in understanding physical phenomena – we often don’t “know” what the machines are “thinking”. For application to aviation, where not only is safety critical but also managing the public perception of risk, this poses a non-trivial psychological hurdle.

Now we get to ethical problems… many of which have already been raised in the case of autonomous cars (e.g. what to do when there is a need to decide between conflicting goals, especially when collateral damage is expected). Can AIs quickly assess the risk to (and both material/non-material value of) civilian life/property in determining an optimal location to perform a crash-landing? And if forced to make a “choice”, who is ultimately responsible? The issue with aviation, as opposed to autonomous cars, is that even though the industry is far more regulated and chances of accidents are far lower, the damage done on a per-accident basis is far greater, as is the social/psychological impact.

The way I see it, safety-critical decisions can often be broken into two categories: predictive actions and reactive actions. Automation can help greatly with predictive actions, especially when systems are complex, and need to be checked frequently.

However, humans are better at reactive actions – diagnosing problems/adapting to issues that have no precedent. Essentially, humans are (as it stands) better at dealing with ambiguity and doing guesswork based on a logical process. It’s no wonder that pilot candidate interviews have questions that test a candidate’s reasoning process and ability to think on his/her feet.

At the end of the day, aviation is a human enterprise that has its roots in hospitality (although, sad to say, this may be changing). The ability of a pilot to empathize and understand the concerns of stakeholders, and the immediacy with which he/she makes decisions is ultimately comforting.

I’m not saying that autonomous airliners will never happen. I just think that it is unwise to underestimate the scale of the challenge.
 
entdoc
Posts: 89
Joined: Tue Jul 26, 2016 11:15 am

Re: Airbus completes testing of self flying jet

Sun Aug 02, 2020 10:41 am

how many computer and sensor systems will there be to assure redundancy in case of failure?
two, three, five? all add weight to the plane maybe even more than a second pilot.
and if both systems fail - a human on board might be able with great difficulty to hand fly. the dead computer probably wont be able to
 
olle
Posts: 2278
Joined: Tue Feb 06, 2007 3:38 am

Re: Airbus completes testing of self flying jet

Sun Aug 02, 2020 11:25 am

I just have a sensation that Airbus plans something for the cargo market.

Will Boeing be able install something similar on 767 if 330F become a single pilot operation?

How much money does it save oer year to run a 330F singlepilot?
 
TheWorm123
Posts: 256
Joined: Sun Oct 06, 2019 8:29 pm

Re: Airbus completes testing of self flying jet

Sun Aug 02, 2020 12:19 pm

airzona11 wrote:
flybynight wrote:
Waterbomber2 wrote:
If they can make cars drive autonomously, why wouldn't they be able to make aircraft fly autonomously? Less obstacles, less improvisation required.

Pilots still have to save the day almost everyday with current tech aircraft but checklists can be run by computers too and faster.
It all is a matter of how many scenarios you can program into the computer, so what OEM need to do is to collect data from as many flights as possible, and program all the scenario's so that issues can be resolved autonomously.

An onboard pilot serves as back-up, but the airplane should also be able to be flown remotely.

Pilots hate to talk about this, but the fact is that it only takes one mistake, one bad judgement, and your children and your passengers' children may never see their mommy or daddy ever again.
This is now a rare occurrence, but if we can do even better, why not?
We are only few years away from launching to Mars with hyper-automated machines, so there'll be jobs for pilots, and they'll be more meaningful and satisfying, no need to be protective.
Let's embrace progress and focus on bigger things.


Well first of all, autonomous cars aren't doing as well as originally thought. A lot factors have come up to put a damper on it.

Humans can predict based on data that computer can never factor in. Could a computer "think" that child on the side of road looks like he/she might jump into the road based on experience and the way the child is behaving?
Also if we go to one pilot, what if he/she becomes sick?
The less pilots fly the plane, the less experience they also rack up. I want my pilot sharp and ready to handle situations. Not as a backup to IBM.


With connectivity, an incapacitated pilot can mean remote take over and descent/land at closest alternative airport. That is within the realm of today's technology. As for the analogy of predicting what the child will do... a computer can make assessments on movement on all sides of the car, from farther distances away than humans. Sure if the driver is solely staring at and aware of that one child, they can react, but automatic breaking for example is a technology that has successfully rolled out to most new model cars to prevent what humans might miss or slowly react to. Insert adding more automation into the cockpit and making flying safer.

But what is the backup for that? Satellites connections can be dodgy and radio waves are crap over the mid Pacific/Indian ocean regions, connectivity issues are still a major problem and wireless network latency issues will also forever be a major problem.
B752 B753 A332 A321 B738
 
User avatar
cirrusdragoon
Topic Author
Posts: 188
Joined: Thu Feb 22, 2018 6:42 pm

Re: Airbus completes testing of self flying jet

Sun Aug 02, 2020 1:07 pm

A great article. I truly believe there will be a singularity moment when yes AI will be as smart as a human brain and exceed it. The technology is accelerating every day. I too agree that there would have to be a fail safe solution , as in a relief pilot onboard the aircraft for the unlikely scenarios.

https://m.economictimes.com/tech/ites/t ... 198656.cms

“AI-driven advancements will see machines mimicking humans and getting better than humans in some tasks. Futurists like Ray Kurzweil say robots will achieve “singularity” — that point when machines become smarter than human beings. This will lead to the development of Artificial General Intelligence (another term for human level intelligence). Kurzweil is betting on machines matching human intelligence by 2029.”
 
olle
Posts: 2278
Joined: Tue Feb 06, 2007 3:38 am

Re: Airbus completes testing of self flying jet

Sun Aug 02, 2020 1:08 pm

<Self driving vehicles will explode the useage for transport sector like heavy trucks, buses taxi etc. In this area the vehicle is used around the clock or close around the clock. Many trucks only stops for chauffeurs sleep or unload. We are soon in a situation where the truckdrivers can set the truck on self driving mode and sleep while the truck is driving. This would increase the utilization of often very expensive vehicles up to 20%. Huge capital savings beside salaries.


This is about to happen right now not in the long distance future.
 
WeatherPilot
Posts: 557
Joined: Sun Jun 25, 2017 1:51 am

Re: Airbus completes testing of self flying jet

Sun Aug 02, 2020 2:11 pm

The problem with computers and flying is they lack the value of self-preservation. The computer doesn’t know what it means to die. It will just compute the most favorable scenario with the information it has available. That’s why pilots will always be necessary, the computer can’t “sense” the outcome of what it will never understand.
 
olle
Posts: 2278
Joined: Tue Feb 06, 2007 3:38 am

Re: Airbus completes testing of self flying jet

Sun Aug 02, 2020 2:30 pm

I think we discuss different things here. I assume that in aviation we talk about going from 2 pilots to one pilot.

In traffic we move to selfdriving on highways. No tired, drunk, suicidal drivers today stands for a major part of serious accidents.

Perhaps number of serious accidents will go down 50% in traffic.

In aviation starting with cargo what is the problem?
 
User avatar
SheikhDjibouti
Posts: 2239
Joined: Sat Sep 30, 2017 4:59 pm

Re: Airbus completes testing of self flying jet

Sun Aug 02, 2020 9:49 pm

WeatherPilot wrote:
The problem with computers and flying is they lack the value of self-preservation.
As compared to say... Korean (& other) co-pilots not daring to speak out against the Captain's actions, because it is not in their culture to question senior positions, even though they are imminently going to die.

Or a Pakistan Airlines Captain performing a go-around after dragging the engines along the runway. Because dying is so much better than losing "face".

Funnily enough, it's usually the other way around; pilots eschewing the correct decision (a go-around) in favor of a botched landing too far down the runway. Who wants to admit they got the approach all wrong?

At least a computer doesn't suffer from pride or vanity.

The computer doesn’t know what it means to die. It will just compute the most favorable scenario with the information it has available.
Sounds good to me! :bigthumbsup:
Of course I am relying on the programmer realizing that landing right way up in a cornfield is (usually) better than upside down on the airport runway.

That’s why pilots will always be necessary, the computer can’t “sense” the outcome of what it will never understand.
Yeah, let's give it up for pilot's and their "senses", such as Atlas Air 3591 First Officer, who "sensed" they were about to stall, and pitched his 767 straight down into the ground. :roll:
Nothing to see here; move along please.
 
airzona11
Posts: 1756
Joined: Wed Dec 17, 2014 5:44 am

Re: Airbus completes testing of self flying jet

Sun Aug 02, 2020 10:20 pm

TheWorm123 wrote:
airzona11 wrote:
flybynight wrote:

Well first of all, autonomous cars aren't doing as well as originally thought. A lot factors have come up to put a damper on it.

Humans can predict based on data that computer can never factor in. Could a computer "think" that child on the side of road looks like he/she might jump into the road based on experience and the way the child is behaving?
Also if we go to one pilot, what if he/she becomes sick?
The less pilots fly the plane, the less experience they also rack up. I want my pilot sharp and ready to handle situations. Not as a backup to IBM.


With connectivity, an incapacitated pilot can mean remote take over and descent/land at closest alternative airport. That is within the realm of today's technology. As for the analogy of predicting what the child will do... a computer can make assessments on movement on all sides of the car, from farther distances away than humans. Sure if the driver is solely staring at and aware of that one child, they can react, but automatic breaking for example is a technology that has successfully rolled out to most new model cars to prevent what humans might miss or slowly react to. Insert adding more automation into the cockpit and making flying safer.

But what is the backup for that? Satellites connections can be dodgy and radio waves are crap over the mid Pacific/Indian ocean regions, connectivity issues are still a major problem and wireless network latency issues will also forever be a major problem.


You can never eliminate 100% of the risk, arguing that 2 humans are the only way to get to peak risk mitigation is counter to path innovation leads society. But in this example, on the edge of envelop ETOPS, fine, keep 2 pilots, start with non-ETOPs type flights. Seems like that would address the risk in the scenario and plausible as ETOPS flights already operate under different conditions.
 
Kilopond
Posts: 561
Joined: Wed Jun 15, 2016 10:08 am

Re: Airbus completes testing of self flying jet

Sun Aug 02, 2020 10:53 pm

Just a side note. The Earth`s magnetosphere is weakening, the upcoming solar cycle 25 is likely to show severe irregularites. This means the future usability of any GNSS (with GPS being one sub-variant of them) is questionable.
 
kalvado
Posts: 2863
Joined: Wed Mar 01, 2006 4:29 am

Re: Airbus completes testing of self flying jet

Sun Aug 02, 2020 11:03 pm

Kilopond wrote:
Just a side note. The Earth`s magnetosphere is weakening, the upcoming solar cycle 25 is likely to show severe irregularites. This means the future usability of any GNSS (with GPS being one sub-variant of them) is questionable.

Why? Satellite navigation doesn't rely on magnetic field, just propagation of radio waves. And I don't see the reason weaker field would make things worse.
As for Van Allen belts - earth synchroneous satellites are already at worst possible altitude, and they last for a while. Navigation orbits are generally lower.
 
TTailedTiger
Posts: 2543
Joined: Sun Aug 26, 2018 5:19 am

Re: Airbus completes testing of self flying jet

Sun Aug 02, 2020 11:07 pm

BA777FO wrote:
FluidFlow wrote:
BA777FO wrote:

How will the computer know when to carry out an unannunciated checklist? Smoke, fire and fumes checklist? How will it recognise airspeed unreliable and determine which one, if any, are reading correctly? How will it determine whether it has a fuel leak or a dodgy guage? Recognise Ice Crystal Icing? Or volcanic ash? How will it determine whether it has had an engine fail or severe damage?

I have no problem talking about it but if a computer cannot sense the above, how can it deal with them?


Smoke, fire and fumes checklist? Ho do you figure out there is fire in the engine or smoke/fumes in the hold?
The aircraft tells you, it already knows it before you do. The checklist? Saved on the computer/tablet you use to execute it. It is already there

How will it recognise airspeed unreliable and determine which one, if any, are reading correctly?
How do you figure out that airspeed is not correct? You read it from the displays in front of you, the aircraft already knows it, same with the checklist, it is just not programmed to execute it.

I could go on with the other mentioned things but as you already know all the information you get are actually from the aircraft and as you solve them by inputting specific actions into the aircraft you are just a middle man that can be replaced by the computer, faster and more reliable.

There are actually scenarios where the computer has its limits but interestingly none of the above mentioned are said scenarios.


You've never flown a commercial airliner, have you? I don't say that to be rude but your condescending tone doesn't fit your ignorance.

Airspeed unreliable? I just read the airspeed? That'd kill me. Which one do I choose when the captain's PFD is telling me we're overspeeding and the FO's PFD is telling me we're in a stall? Which one will the new autonomous aircraft choose?

On the issue of smoke, the aircract can detect it in the lavatories or crew rest areas but can't detect nor determine the source of smoke from air conditioning or IFE. It cannot detect Ice Crystal Icing, nor volcanic ash.

This is the point of unannunciated checklists that's you've missed - they are unannunciated because the aircraft cannot detect, nor diagnose the problem. It's up to the crew to diagnose the problem and select the appropriate checklist(s). I've not even touched on issues where a binary computer cannot make decisions about delays, usable fuel and suitable alternatives given weather conditions.

Please don't tell me the aircraft can do things it can't. 10,000 hours of flying 737s and 777s has given me the experience to determine otherwise!



Well said. I'm appalled by some of the posts here. You can't call yourself an aviation enthusiast and cheer on eliminating pilots.
 
TheWorm123
Posts: 256
Joined: Sun Oct 06, 2019 8:29 pm

Re: Airbus completes testing of self flying jet

Sun Aug 02, 2020 11:18 pm

airzona11 wrote:
TheWorm123 wrote:
airzona11 wrote:

With connectivity, an incapacitated pilot can mean remote take over and descent/land at closest alternative airport. That is within the realm of today's technology. As for the analogy of predicting what the child will do... a computer can make assessments on movement on all sides of the car, from farther distances away than humans. Sure if the driver is solely staring at and aware of that one child, they can react, but automatic breaking for example is a technology that has successfully rolled out to most new model cars to prevent what humans might miss or slowly react to. Insert adding more automation into the cockpit and making flying safer.

But what is the backup for that? Satellites connections can be dodgy and radio waves are crap over the mid Pacific/Indian ocean regions, connectivity issues are still a major problem and wireless network latency issues will also forever be a major problem.


You can never eliminate 100% of the risk, arguing that 2 humans are the only way to get to peak risk mitigation is counter to path innovation leads society. But in this example, on the edge of envelop ETOPS, fine, keep 2 pilots, start with non-ETOPs type flights. Seems like that would address the risk in the scenario and plausible as ETOPS flights already operate under different conditions.

But if this innovation just for the sake of innovation or a necessary evolution?

Myself I think it’s the former, an AI will never be able to independently make decisions in a circumstance it hasn’t ever seen before. Machine learning algorithms can only apply what they’ve already ‘learnt’ and can only do this in a specific format. I don’t think AI will ever be able to think like a brain for that reason, I say that as a Computer Science graduate.

Maybe when quantum computing becomes more prevalent that might change. We do have some quantum computing capabilities now but it would be decades before you’d see them in aviation.
B752 B753 A332 A321 B738
 
User avatar
cirrusdragoon
Topic Author
Posts: 188
Joined: Thu Feb 22, 2018 6:42 pm

Re: Airbus completes testing of self flying jet

Mon Aug 03, 2020 1:02 am

TheWorm123 wrote:
airzona11 wrote:
TheWorm123 wrote:
But what is the backup for that? Satellites connections can be dodgy and radio waves are crap over the mid Pacific/Indian ocean regions, connectivity issues are still a major problem and wireless network latency issues will also forever be a major problem.


You can never eliminate 100% of the risk, arguing that 2 humans are the only way to get to peak risk mitigation is counter to path innovation leads society. But in this example, on the edge of envelop ETOPS, fine, keep 2 pilots, start with non-ETOPs type flights. Seems like that would address the risk in the scenario and plausible as ETOPS flights already operate under different conditions.

But if this innovation just for the sake of innovation or a necessary evolution?

Myself I think it’s the former, an AI will never be able to independently make decisions in a circumstance it hasn’t ever seen before. Machine learning algorithms can only apply what they’ve already ‘learnt’ and can only do this in a specific format. I don’t think AI will ever be able to think like a brain for that reason, I say that as a Computer Science graduate.

Maybe when quantum computing becomes more prevalent that might change. We do have some quantum computing capabilities now but it would be decades before you’d see them in aviation.


So we mustn’t say never. I agree, what we know about artificial intelligence now will be very different in 60 years into the future. Computer science is constantly evolving. What we couldn’t possible fathom in the 1800’s we are doing now today. It will take time but progress will be made where technology further advances our species.
 
TheWorm123
Posts: 256
Joined: Sun Oct 06, 2019 8:29 pm

Re: Airbus completes testing of self flying jet

Mon Aug 03, 2020 1:59 am

cirrusdragoon wrote:
TheWorm123 wrote:
airzona11 wrote:

You can never eliminate 100% of the risk, arguing that 2 humans are the only way to get to peak risk mitigation is counter to path innovation leads society. But in this example, on the edge of envelop ETOPS, fine, keep 2 pilots, start with non-ETOPs type flights. Seems like that would address the risk in the scenario and plausible as ETOPS flights already operate under different conditions.

But if this innovation just for the sake of innovation or a necessary evolution?

Myself I think it’s the former, an AI will never be able to independently make decisions in a circumstance it hasn’t ever seen before. Machine learning algorithms can only apply what they’ve already ‘learnt’ and can only do this in a specific format. I don’t think AI will ever be able to think like a brain for that reason, I say that as a Computer Science graduate.

Maybe when quantum computing becomes more prevalent that might change. We do have some quantum computing capabilities now but it would be decades before you’d see them in aviation.


So we mustn’t say never. I agree, what we know about artificial intelligence now will be very different in 60 years into the future. Computer science is constantly evolving. What we couldn’t possible fathom in the 1800’s we are doing now today. It will take time but progress will be made where technology further advances our species.


That’s a very good point. I probably shouldn’t have used the word ‘never’ though, after all there are no absolutes in science whatsoever, which is why it’s all based in probability.
B752 B753 A332 A321 B738
 
rbretas
Posts: 63
Joined: Mon Aug 22, 2016 10:21 am

Re: Airbus completes testing of self flying jet

Mon Aug 03, 2020 4:53 am

flybynight wrote:
Humans can predict based on data that computer can never factor in. Could a computer "think" that child on the side of road looks like he/she might jump into the road based on experience and the way the child is behaving?


This resumes well the topic of automation and how most people have no idea about how it works or how much it has changed in the past 10 years.

Computers actually can very easily make that kind of prediction based on experience just as humans do. Even without being a computer science major I am able to use computer vision and train neural networks at work daily exactly for predicting behavior. And the AI does a much better job than me or my colleagues (who are specialists in it) could.
 
strigon07
Posts: 2
Joined: Tue Mar 17, 2020 8:56 am

Re: Airbus completes testing of self flying jet

Mon Aug 03, 2020 6:30 am

TheWorm123 wrote:
But if this innovation just for the sake of innovation or a necessary evolution?


I agree with your sentiment - again, not trying to be too much of a wet blanket here, but we can't always assume that automation will definitely yield benefits. An understanding of maturity of the technology, cost of implementation, and competitiveness is required. What Airbus is currently doing are just baby steps in the grand scheme of things. Put another way: take-offs and landings cover just the begining phase(s) of flight school. There's a lot more that needs to be integrated into a fully autonomous system.

TheWorm123 wrote:
Myself I think it’s the former, an AI will never be able to independently make decisions in a circumstance it hasn’t ever seen before. Machine learning algorithms can only apply what they’ve already ‘learnt’ and can only do this in a specific format. I don’t think AI will ever be able to think like a brain for that reason, I say that as a Computer Science graduate.

Maybe when quantum computing becomes more prevalent that might change. We do have some quantum computing capabilities now but it would be decades before you’d see them in aviation.


Side note: as a physics graduate, agreed on the decades assessment on quantum computing too. Cold atoms are very, very sensitive to thermal/environmental noise, and optical systems can take weeks to months just to calibrate. Now imagine putting those into an airliner that has to deal with turbulence and cosmic radiation... not to mention the weight from all the insulation that will likely be needed. Hopefully there's progress in scaling quantum technologies to room temperature.
 
User avatar
cirrusdragoon
Topic Author
Posts: 188
Joined: Thu Feb 22, 2018 6:42 pm

Re: Airbus completes testing of self flying jet

Mon Aug 03, 2020 5:20 pm

I quite enjoy the views of Erik Brynjolfsson, director of the MIT Initiative on the Digital Economy and author of “Machine, Platform, Crowd: Harnessing Our Digital Future”

“... AI and related technologies have already achieved superhuman performance in many areas, and there is little doubt that their capabilities will improve, probably very significantly, by 2030. … I think it is more likely than not that we will use this power to make the world a better place. For instance, we can virtually eliminate global poverty, massively reduce disease and provide better education to almost everyone on the planet. That said, AI and ML [machine learning] can also be used to increasingly concentrate wealth and power, leaving many people behind, and to create even more horrifying weapons. Neither outcome is inevitable, so the right question is not ‘What will happen?’ but ‘What will we choose to do?’ We need to work aggressively to make sure technology matches our values. This can and must be done at all levels, from government, to business, to academia, and to individual choices.” https://www.pewresearch.org/internet/20 ... of-humans/

So it will definitely remain to be seen how the aviation economy shifts its values in terms of profits , efficiency, safety amongst other factors not mentioned for sure.
 
bhill
Posts: 1837
Joined: Thu Sep 13, 2001 8:28 am

Re: Airbus completes testing of self flying jet

Mon Aug 03, 2020 6:25 pm

TObound wrote:
The comments here are naive. The amount of people who think automation can't replace aircrew sounds like every other sector and job before they were automated out.

Sure it may not happen soon. And it will be a process. But it is coming. The speed with which AI can analyze and make decisions, along with the cost of pilots, and the relatively lower workload at cruise (ie. Lower productivity) is exactly why their jobs are targets. Even eliminating a relief pilot or one in the cockpit is a huge savings.

Want to know who has already been through the process? Military combat aircraft. Look up how checklists work in an emergency on the F-22. And those have to be designed for combat damage. Automation and improved sensors has enabled single seat cockpits for air forces and made those aircraft more capable over the years. There is no reason the same can't be done in the transport sector.


You are the one being naive. Military service members KNOW they may be killed in their duties, there is an "acceptable" loss percentage calculated in for any combat operation..commanders desire it be near zero, but some may happen. Civilians will not tolerate ANY loss getting onto a jetliner. So, how would you calculate an "acceptable" failure rate for AI on a jetliner that will cause death? There is a reason a human will ALWAYS be in the active loop of civilian transport...even if remotely like a drone. 100% autonomous AI will never happen in civilian passenger aviation.
Carpe Pices
 
144modeller
Posts: 39
Joined: Tue Sep 27, 2016 9:52 pm

Re: Airbus completes testing of self flying jet

Mon Aug 03, 2020 8:06 pm

SheikhDjibouti.
With all your expertise, I'm surprised you don't rule the world.

At this moment, there are thousands of people claiming that the Coronavirus is a con.
 
User avatar
SheikhDjibouti
Posts: 2239
Joined: Sat Sep 30, 2017 4:59 pm

Re: Airbus completes testing of self flying jet

Mon Aug 03, 2020 9:02 pm

144modeller wrote:
SheikhDjibouti.
With all your expertise, I'm surprised you don't rule the world.
I suggest you acquaint yourself with the rules on personal attacks.

I'm no angel in this respect myself, but at least I wrap it up with some factual presentation relevant to the discussion, which indirectly proves what an idiot the other guy is. (That's a general comment; not aimed at you in any way)

But you didn't even try. No worries because that's ok; instead of reporting you, I prefer to leave your post as evidence for others to enjoy. :wave:


144modeller wrote:
At this moment, there are thousands of people claiming that the Coronavirus is a con.
And this is relevant in what way exactly?
I'm open to a persuasive argument from you, but tbh I don't even know where you'd start. #quite_puzzled
Nothing to see here; move along please.
 
IADFCO
Posts: 190
Joined: Sun May 22, 2016 4:20 pm

Re: Airbus completes testing of self flying jet

Mon Aug 03, 2020 10:49 pm

Most machine learning algorithms are glorified interpolation algorithms. As such, they perform well within the data (e.g., the flight conditions) used for the training, but how well they can extrapolate is usually a big question mark. Neural Networks are polynomial interpolation on steroids, "deep learning" NN are simply NN with more terms. "Nearest Neighbor" algorithms also are good within the training data and don't work well outside. Ditto for clustering algorithms and principal component analysis. Also, all these algorithms are often subject to the so-called "curse of dimensionality", which means that as the number of parameters that one wants the AI system to consider increases, the computational effort increases much much faster.

You don't need machine learning to train an AI engine, but then you have to come up with the rules, and make sure you cover all possibilities for the holes in the cheese to line up.

Bottom line, IMHO, for limited tasks these systems can be very useful, but a truly "self flying jet" is far beyond the state of the art. Military applications, where money is no object, people on board (if present) accept that they can be killed, and often there are ejection seats, are easier.
 
frmrCapCadet
Posts: 4262
Joined: Thu May 29, 2008 8:24 pm

Re: Airbus completes testing of self flying jet

Tue Aug 04, 2020 12:36 am

Zero loss is not and never will be possible. Reality res safety is statistical.
Buffet: the airline business...has eaten up capital...like..no other (business)
 
TObound
Posts: 781
Joined: Mon May 27, 2019 12:54 am

Re: Airbus completes testing of self flying jet

Wed Aug 05, 2020 5:07 pm

bhill wrote:
TObound wrote:
The comments here are naive. The amount of people who think automation can't replace aircrew sounds like every other sector and job before they were automated out.

Sure it may not happen soon. And it will be a process. But it is coming. The speed with which AI can analyze and make decisions, along with the cost of pilots, and the relatively lower workload at cruise (ie. Lower productivity) is exactly why their jobs are targets. Even eliminating a relief pilot or one in the cockpit is a huge savings.

Want to know who has already been through the process? Military combat aircraft. Look up how checklists work in an emergency on the F-22. And those have to be designed for combat damage. Automation and improved sensors has enabled single seat cockpits for air forces and made those aircraft more capable over the years. There is no reason the same can't be done in the transport sector.


You are the one being naive. Military service members KNOW they may be killed in their duties, there is an "acceptable" loss percentage calculated in for any combat operation..commanders desire it be near zero, but some may happen. Civilians will not tolerate ANY loss getting onto a jetliner. So, how would you calculate an "acceptable" failure rate for AI on a jetliner that will cause death? There is a reason a human will ALWAYS be in the active loop of civilian transport...even if remotely like a drone. 100% autonomous AI will never happen in civilian passenger aviation.


Have you served? Because this is ignorant. There's no acceptable loss in the military for just flying around normally. There is risk acceptance during tactical training or combat.

Popular Searches On Airliners.net

Top Photos of Last:   24 Hours  •  48 Hours  •  7 Days  •  30 Days  •  180 Days  •  365 Days  •  All Time

Military Aircraft Every type from fighters to helicopters from air forces around the globe

Classic Airliners Props and jets from the good old days

Flight Decks Views from inside the cockpit

Aircraft Cabins Passenger cabin shots showing seat arrangements as well as cargo aircraft interior

Cargo Aircraft Pictures of great freighter aircraft

Government Aircraft Aircraft flying government officials

Helicopters Our large helicopter section. Both military and civil versions

Blimps / Airships Everything from the Goodyear blimp to the Zeppelin

Night Photos Beautiful shots taken while the sun is below the horizon

Accidents Accident, incident and crash related photos

Air to Air Photos taken by airborne photographers of airborne aircraft

Special Paint Schemes Aircraft painted in beautiful and original liveries

Airport Overviews Airport overviews from the air or ground

Tails and Winglets Tail and Winglet closeups with beautiful airline logos