planemaker
Topic Author
Posts: 5411
Joined: Mon Aug 25, 2003 12:53 pm

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 6:19 am

Quoting maxpower1954 (Reply 47):
And I'm not talking about big airplane versus little one; it's the complexity of the operation and the constant decision making that goes with the territory.

I'm extremely familiar with the complexity.To use Light Saber's term, you are looking at if via "a rear view mirror" because the current technology in aviation is far from leading edge. Even though Airbus improved the A321, it still nevertheless got it's birth, in essence, some 30 years ago... which in tech terms is a couple of centuries ago.  
Quoting maxpower1954 (Reply 47):
My experience and judgement isn't replaceable for the forseeble future meaning 20 -30 years at least down the road.

It will be... cognitive computing. Furthermore, while it took years to acquire valuable piloting experience and judgement (which varies significantly from pilot to pilot), it is instantly replicable in a computer. For example, in a couple of years IBM's Watson will pass the Medical Licencing Exam and all "Watson's" will have the same capability.

Quoting maxpower1954 (Reply 47):
I Googled Missy Cummings, and didn't find her wanting to replace the pilot in a commercial operation.

That is not her area of work. However, she is on record (print and video) that computers fly aircraft better than pilots (she claims that she has received a ton of hate mail because of her statements) and that you don't pilots in the cockpit.

Quoting FlyHossD (Reply 48):
Is there some reason why these proposed regulations (see page 33, 34, etc.) wouldn't apply to a "PrimeAir" type of operation:

Yes, starting with they are written for UAS ops in commercial airspace co-mingling with commercial air traffic. As pointed out earlier, the FAA's real concern is a UAS collision with an airliner.

Quoting lightsaber (Reply 49):
What is laughed at as not achieved in one year is easily met two years later.

Just as what happened with DARPA's Grand Challenge and then their Grand Urban Challenge.

Quoting lightsaber (Reply 49):
Imagine what could be designed with Fin-Fet transistors (common by 14nm production).

I know, I know!   

And they will be relatively cheap..
Nationalism is an infantile disease. It is the measles of mankind. - A. Einstein
 
tim73
Posts: 93
Joined: Sun Nov 17, 2013 9:03 pm

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 9:18 am

Flying is hard for humans to learn because humans are pretty lousy at multitasking. Driving is also quite hard to learn but the required level of multitasking is lower.

For computers, flying is easier because it is mostly just finding out the required flight envelope at any given situation, based on the laws of physics. You know what the limits are even inside a big storm. Tolerance limits are also higher than with driving, you do not need to land within centimeters with commercial airliners.

You also got your nice, predetermined glide paths with every airport out there and radio beacons helping all the way to airfield and even set of lamps showing whether you are still in the ideal path. That is practically ideal for computers to do!

Driving is much harder for computers because it requires a lot more subjective decisions based on things happening outside of car. How to indentify a road repair site all of sudden, how to avoid a hole in the road, how to identify pedestrians etc etc.

The number of test hours required is at least 100 times higher than for automated planes. That is why Google and others are out there driving all the time to test their cars. A350 is going to need mere 2500 flight hours for certication with quite a lot of automation. I'd bet it could even autoland already. With automated cars it is going to need at least 250 000 hours of testing out there!
 
Pihero
Posts: 4318
Joined: Mon Jan 31, 2005 5:11 am

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 12:08 pm

Just gave this thread a quick sight.
You are just repeating your mantra : automatons to-morrow and pilots out.
Gets boring...

You haven't begin to give an answer to the main questions :
- What is intelligence ? The question is especially important as IT talks about logic, defined by the Oxford Dictionary as "a set of principles underlying the arrangements of elements in a computer... so as to perform a specified task"... and if you don't have the first clue as to what the mutiple tasks are, your programming is just a load of smelly stuff.
- What are the protections when things go awry on-board or around an automated vehicle ? (and , no your remote operator isn't the answer for reasons of comm lag and encription safety. To say the contrary is disingenuous or downright dishonest.).

SESAR ( the acronym is for *Single Europeanb Sky ATC - or ATM - Research* ) has so far a development budget of 2.1 billion € and plans to see applicable results not before 2030... and it's just a modernisation of the airspace control / traffic control / departue, route and arrival management-. Nothing is said about drone or automated aircraft in the plans... So, with actual technology and a projectionj of achievable new tec, just smoothing up the airspace for its application is outside your pipe dream of flying automatons.
Someone is certainly out of his head, here.
I wonder who ...   

Quoting planemaker (Reply 40):
Professor Missy Cummings at MIT (ex-Navy F/A-18 pilot

Worth repeating Ms Cummings declaration :

Quoting maxpower1954 (Reply 47):
I’m not trying to take automation and replace humans. My research is how to develop collaborative systems so that automation is enabling people to do their jobs better."

Not very surprising when one of her works is titled as :"Levaraging Human-Computer collaboration for Decision-Making in Complex Systems".
I'm trying to find that book.

I propose some reading on this subject :

1/-" Aviation Automation : The Search For a Human-Centered Approach " by Charles Billings
and
2/-" Automation Airmanship" by Chris Lutat and Ryan Swah
Contrail designer
 
FlyPNS1
Posts: 5480
Joined: Wed Dec 01, 1999 7:12 am

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 1:22 pm

Quoting lightsaber (Reply 49):
So you're not participating in NASA's April (2014) challenge?
http://www.nasa.gov/press/2013/septe...pi-unmanned-aircraft/#.UsTkJIuzKpg

This will be like the DARPA robotic challenges. What is laughed at as not achieved in one year is easily met two years later.

The technology is being worked. Heck, college kids participate a la the SAE races:
http://en.wikipedia.org/wiki/International_Aerial_Robotics_Competition

The milestones are being met far faster than pundits have predicted:"Three years later in 1995 a team from Stanford University was able to acquire a single disk and move it from one side of the arena to the other in a fully autonomous flight%u2014half a decade earlier than some pundits had predicted."

Every one of their scenarios was considered impossible, but eventually college teams meat and beat the challenges.

There's a big difference between a bunch of college kids tinkering around in a lab and the real world. I don't think anyone is really questioning the technology, but it's the logistics, regulations and economics where we've barely scratched the surface. And those areas tend to move much more slowly than technology.

Quoting planemaker (Reply 50):
Yes, starting with they are written for UAS ops in commercial airspace co-mingling with commercial air traffic. As pointed out earlier, the FAA's real concern is a UAS collision with an airliner.

That's one big concern for sure, but with drones the FAA (and the government at large) will have to worry about a whole lot more. For example, drones flying too close to buildings will create a whole new set of policy issues regarding personal privacy and property protection.
 
NAV20
Posts: 8453
Joined: Thu Nov 27, 2003 3:25 pm

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 1:51 pm

Just an idle thought really - but it occurs to me that driving is basically a matter of two dimensions, whereas flying involves three? And I can say, on my own account, that flying any sort of aeroplane (at the start) was a damn sight scarier than driving a car. Even in perfect visibility.....

Maybe seek to perfect two-dimensional 'automatic driving' first? And wait until that's been proved not to be 'life-threatening' before we move on to three-dimensional 'automatic flying'?
"Once you have flown, you will walk the earth with your eyes turned skywards.." - Leonardo da Vinci
 
tim73
Posts: 93
Joined: Sun Nov 17, 2013 9:03 pm

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 2:40 pm

Quoting Nav20 (Reply 54):
Maybe seek to perfect two-dimensional 'automatic driving' first? And wait until that's been proved not to be 'life-threatening' before we move on to three-dimensional 'automatic flying'?

Actually, car related software is the most complex one.

Space shuttle codebase: 40 000 lines of code.
US military drone, control: 3.5 million
Firefox browser: 9 million
Windows NT 1996: 12 million
Boeing 787, avionics only: 6.5 million
Boeing 787, total: 14 million lines
Linux 3.1: 15 million
F-35 Fighter: 24 million
Windows 7: 39 million
Large Hadron Collider: 50 million
and then....average modern high end car: 100 million!

http://www.informationisbeautiful.ne...ualizations/million-lines-of-code/

What is difficult for humans, does not mean it is difficult to program with computers. What is easy for humans, could be instead very difficult to program like driving with constant flood of information from surroundings.

[Edited 2014-01-02 06:50:32]
 
User avatar
lightsaber
Moderator
Posts: 18672
Joined: Wed Jan 19, 2005 10:55 pm

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 3:58 pm

Quoting planemaker (Reply 50):
Just as what happened with DARPA's Grand Challenge and then their Grand Urban Challenge.

Exactly! And all before Fin-Fet transistors which will allow for much more software to do its job.

Quoting FlyPNS1 (Reply 53):
There's a big difference between a bunch of college kids tinkering around in a lab and the real world.

My career is bridging that transition. I'm seeing the hundreds of millions (if not billions) being spent to make it happen. Due to high fuel prices, the economics will push the transition. We won't see cheap oil again as every $5/bbl drop in oil below about $90/bbl removes 1 million bbd off the global market (economics to achieve the return on investment). We've never before had such a rapid rise in oil prices.

There will be a delay. But from 2002 to 2008, I saw the field stagnate. Since 2008 I've seen tremendous progress.

Quoting tim73 (Reply 55):
What is difficult for humans, does not mean it is difficult to program with computers.

Certifying autonomous aircraft is so much simpler than with humans. One can test the branches of the software which speeds things up prior to flight.


Lightsaber
IM messages to mods on warnings and bans will be ignored and nasty ones will result in a ban.
 
wowpeter
Posts: 86
Joined: Fri Oct 20, 2006 1:15 am

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 5:41 pm

First of all... sorry for my bad grammar...

Quoting Rara (Reply 18):
Yes, but when does that ever happen? When did lightning last knock all electronics out, but the aircraft was still recovered by a human pilot? That's practically a non-possibility.
Quoting tim73 (Reply 19):

#1: Engineers today are very good at shielding electronics from lightning. Anything short of nuclear EMP strike nearby can be dealt with (fighter planes are even shielded against those btw). Lighting protection circuit reaction time is calculated in NANOSECONDS today. If it were somehow possible to generate so powerful lightning strike, it would knock out much more than just the comms. Possibly blow up the plane itself right away. But then again you would be flying in planet Jupiter atmosphere

I have an engineering background and I am also a professional pilot... the so call non-possibility is very possible... We can all think we will built the most fail proofed system, and shield everything like crazy... but what happen if the maintainance people mess up and didn't ground the electronic properly after maintainance? Things like that can happen right? What happen if there is a batch of defective equipment during manufacturing? This happens so often it is not even funny... So what happen if all those fails you and something like this did indeed happen? It is a possibility, something that we need to do to think outside the box. We can't just say not possible and be done with it... a lot of people say, Air France is not possible, or the Qantas A380 incident type of damage is not possible, but it still happens ever so often...

Finally, I can assure you, most plane that you are flying now (maybe with a few rare exception), you can still have minimal flying control without any electronics... so there is still a chance a genius pilot will be able to land safely... or a average pilot will be able to crash land it with some survivor...

Quoting Rara (Reply 18):
The Air France plane was doomed, with two and a half human pilots in the cockpit. Chances are very high that the plane wouldn't have crashed, it if had been piloted by an artificial intelligence which would just have followed the procedures.

AF447 is probably the worst argument in favour of human pilots you could come up with.
Quoting tim73 (Reply 19):
#2: Data would have been conflicting but somebody could have said from the "batcave" to fly level with 85 percent power etc etc. Somebody seeing the forest out of the trees...anyway there would be multiple level of comms, even portable satellite phones. If the lightning were able to fry even those, the crew and the passengers would be fried to death too.

What you claim about AF447 being the worst argument is not true and this is not a bad example like you claim... you only think it is bad example because you and I have the benefit of hindsight... we now know what we should do with AF incident... just set 80% power with 2.5 degree pitch up... but if a computer is so smart, why couldn't they built a system smart enough to not let the auto-pilot drop out in the first place? They can't do that because Airbus or Boeing for that matter, doesn't think you will get conflicting data on all 3 systems at the same time... it is not a scenario that they design the system for... hence, if this happen, they will just drop the autopilot and let the pilot deal with it... Once again, this is our think inside the box way of doing things... what do we think is the worse case... but there could always be a worse worse case, but no no no, it can never happen... but things like that keeps happening... QF A380 incident... oh no, that part inside the turbine will never fail, it is a life failure free parts, so we design it to never fail for the life of the parts, so this is why we don't need to shield the core of the engine with kevlar and only need to do it to the Fan crowling... and look what happen... No No No, it can never happen, and then it happen... our thinking inside the box philosophy is what failing us really...

Finally, back to the scenario, a computer system afterall rely on data input... when you have corrupted data input, it is very difficult for a computer system to figure out what went wrong... For example, if I modify the Air France situation a little bit and let say we add some severe airframe icing which causes the 80% power and 2.5 degree becomes a solution that will still crash the plane? And with 80% power and 2.5 degree pitch up, the airplane still continued to stall and fall out of the sky... and let say, the only way for it not to un-stall is to lower the nose to 1 degree pitch down... but no one knows that, the human doesn't know, the computer doesn't know... so what do you think a confuse computer will do? And what will the human pilot be able to do?

The problem that I have is that the people who argue these points and saying computer will replace humans, are still thinking inside the box... but what happen if something did happen outside that box? What will the computer / artificial intelligence be able to do? I know what a human will be able to do / can do...

Quoting Rara (Reply 18):
Again, when did that ever happen? When did a human pilot successfully crash-land a burning plane that an artificial pilot could not have saved?

Let's not forget that with nearly 100.000 scheduled flights every day, 365 days a year, we have an insanely good database of what kind of stuff actually happens to airplanes, and what kind of stuff doesn't. While a situation where a human pilot crashes a burning plane into the ground to save some few lifes could theoretically happen, in practice it doesn't. So that's really no argument against automated planes.

Once again you are thinking inside the box... when did that ever happen? Because it is something that could happen... it just haven't happen yet, or that it has indeed happen many times, but so far we all failed to resolve it... UPS in Dubai, Asiana on the way to Korea, Swireair, South Africa 295, they are the unlucky one... the lucky one, like Singapore Airline in Bangkok... Or the false warning, like Cathay 777 back in Kai Tai... So it has happen... and it is just that when shit hits the fan, no pilot have actually try crash landing it to try to save at least a few peoples life instead of hanging on and kill everyone... but at least where things in pilot training are changing... there is the thought at my airline that if situation is dire enough, ditching or force landing a big jet is something that we should consider in dire situation... we should not only think landing on runway only... sometimes, situation may be dire enough that saving a few lives is better than burn apart and killing everyone...

Quoting tim73 (Reply 19):
#3: Well, Swissair Flight 111 crew could not do anything anyway with onboard fire.

Yes, there are always going to be one out of one million situations where being present would be more desirable or even crucial. But if the overall safety level can be increased with remote pilots, it will be done. Each of those one out ouf one million situations will be also taken into account in the next plane upgrades.

The last generation planes, 777 and 330s, were more automated because pilots were not that trustworthy. Jets are just too slick and fast to be left to be controlled by pilots alone, too easy to make mistakes. Only mostly former fighter pilots are that good (Finnair with unbeatable safety record for example had a lot of them in the 70-90's). Automation today is there to make the job easier and either warn (Boeing, soft limits) or prevent mistakes (Airbus, flight envelope with hard limits).

Pilots in bigger planes are already almost computer deskjockeys, they just do not realize it

For your info, if Swireair were to divert and land with an overweight landing, they would have live, but they elect to dump fuel, wasting precious time... Pilot are now taught not to do that... Just go overweight landing in cargo fire... you only have 15 mins or else you will need to put it in the water or on the ground.

Secondly, I would love to be able to agree with you that pilot in bigger plane are almost computer deskjockey... but unfortunately i can not... for example today... on a fully managed arrival into HKG, there is a restriction at a certain waypoint that the plane needs to be at FL260... and thanks to our lovely computer, it is unable to do it automatically? Why? I donno why the computer couldn't do it... there is nothing out of the ordinary, relatively high speed and fully managed speed, no significant wind changes or wind deviation from the uploaded wind... so why the computer fail to achieve what it is program to achieve and I, as a pilot need to intervene about 10 miles before that way point and manually deploy some speedbrake inorder to meet ATC requirement? Why? The computer did not even ask me to deploy speedbrake like they normally will sometimes, the computer thinks it can do it, but we can see that it is running away from it on the crossing altitude info... Since I, as a human pilot, saw the requirements keep running away from us and the computer did nothing under fully automatic (managed mode), not even ask me to deploy speedbrake to help descent... So if the computer is really that smart, shouldn't it be able to correct itself? No? This is because computer are stupid... human pilots on modern commerical aircraft, inorder to be able to operate them safely, need to be at least mentally ahead of the aircraft to be able to monitor if the computer performing as per it is programmed to do... if not, we as human pilot need to intervene... So right now, as much as I like to think modern commercial aircraft computer is very smart and we are just deskjockey, it is unfortunately not the case... In so many instance, I end up having to intervene so that computer doesn't mess up... And btw, I fly an Airbus 330... and I can tell you, their software is crap... even the most basic thing... Sigh... only if the average people know this...
 
planemaker
Topic Author
Posts: 5411
Joined: Mon Aug 25, 2003 12:53 pm

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 6:38 pm

Quoting lightsaber (Reply 56):
My career is bridging that transition. I'm seeing the hundreds of millions (if not billions) being spent to make it happen.

It is in the BILLIONS. Unfortunately most people don't have an understanding of what is going on worldwide in the field of AI and robotics. Some of the stuff that is being developed in Israel is way ahead of what we have. And it is not just the military that is footing the bill (though they certainly are funding a big chunk of it). As I have posted previously, there is IBM with their advances in cognitive computing and deepQA (and IBM is the only company that people on here might have an inkling about it this field). Some people might know that Amazon is the largest online retailer but are completely oblivious to the fact that they are one of the biggest web services provider with clients including the CIA and NASA and that they own a robotic company. And Google is so much more than a search engine: Boston Dynamics, Meka Robotics, Schaft inc., Holmni, Industrial Perception, Redwood Robotics, Bot & Dolly, and of course, Motorola. And all of the above is just barely scratching the surface. Unfortunately, people just haven't a clue about what is going on in this field much less being able to grasp the implications.

Quoting lightsaber (Reply 49):
This will be like the DARPA robotic challenges.

BTW, at the recent DARPA Robotic Challenge Trial, Google's robots placed 1st, 2nd and and 4th.   

Quoting FlyPNS1 (Reply 53):
That's one big concern for sure, but with drones the FAA (and the government at large) will have to worry about a whole lot more. For example, drones flying too close to buildings will create a whole new set of policy issues regarding personal privacy and property protection.

All of that is easily solved (remember Bab's attempt to stop a helicopter from filming the CA coastline??) but your comment illustrates the narrow vision that the vast majority of people have on the subject. Where the big profits are to be made with UAVs is in sparsely settled or remote areas flying at very low altitude. And that is why there is the legal challenge... it obviously makes no sense to not allow it.

Quoting wowpeter (Reply 57):
We can't just say not possible and be done with it... a lot of people say, Air France is not possible, or the Qantas A380 incident type of damage is not possible, but it still happens ever so often...

To borrow Lightsaber's comment again, you are looking through the "rear view mirror." When was the A330 designed? Even the AHMS on the A380 is old in tech terms.
Nationalism is an infantile disease. It is the measles of mankind. - A. Einstein
 
FlyHossD
Posts: 1975
Joined: Mon Nov 02, 2009 3:45 pm

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 7:47 pm

Quoting planemaker (Reply 50):
Yes, starting with they are written for UAS ops in commercial airspace co-mingling with commercial air traffic. As pointed out earlier, the FAA's real concern is a UAS collision with an airliner.

Nice side-step, I think. In other words, you didn't address my question. If Amazon, etc. start using drones for delivery, the FAA will consider them to be commercial operations. Therefore, they'll be regulated. Low-level drones will interfer with other air traffic, especially helicopters.

Eventually, I do believe that unmanned airliners will exist, but that's probably several decades down the road. During my career, I saw the "impossible" happen in the cockpit more times than I can count.

One more thing, ALL of my old friends are engineers and apparently good ones, at that (they've had very successful careers). To a man, they all exhibit the "engineer's arrogance" that I'm seeing here. That is, the engineers/designers ALWAYS overlook something - it's just a fact of life.

[Edited 2014-01-02 11:50:48]

[Edited 2014-01-02 12:42:49]
My statements do not represent my former employer or my current employer and are my opinions only.
 
FlyPNS1
Posts: 5480
Joined: Wed Dec 01, 1999 7:12 am

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 8:19 pm

Quoting planemaker (Reply 58):
All of that is easily solved (remember Bab's attempt to stop a helicopter from filming the CA coastline??) but your comment illustrates the narrow vision that the vast majority of people have on the subject.

If it was easily solved, we'd already have a solution. You fail to comprehend some of the legal and regulatory complexities and how slow the FAA moves.

Quoting planemaker (Reply 58):
Where the big profits are to be made with UAVs is in sparsely settled or remote areas flying at very low altitude.

Talk about no vision...what you described is really the tip of the iceberg. The bulk of UAV operations will eventually be in urban areas. And that's the big challenge...not a few drones dumping pesticides on a farm. Amazon wants to deliver packages to where people live and most American's live in urban/suburban areas. That's where they will make the big bucks. Delivering packages to a few farmers is nice, but not where the big bucks will be.
 
User avatar
lightsaber
Moderator
Posts: 18672
Joined: Wed Jan 19, 2005 10:55 pm

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 8:31 pm

Quoting planemaker (Reply 58):
It is in the BILLIONS. Unfortunately most people don't have an understanding of what is going on worldwide in the field of AI and robotics. Some of the stuff that is being developed in Israel is way ahead of what we have.

I'm glad I put billions in parenthesis.   I'm going to disagree on Israel being way ahead of our latest UAVs. They're ahead in what they're willing to sell. No doubt about that.

Quoting planemaker (Reply 58):
Unfortunately, people just haven't a clue about what is going on in this field much less being able to grasp the implications.

The implications are huge. Even though I try to follow, the development is happening at an amazing pace with new players surprising everyone.

FWIW, my brother works at a robot vendor and both of us are continually surprised at the rate of progress. His company did a software update that *more than doubled* the robot efficiency.    So much so that their customers couldn't staff up fast enough to do the 'support the robot jobs' to help the robots work at their peak efficiency... Classy problem to have.

Quoting planemaker (Reply 58):
BTW, at the recent DARPA Robotic Challenge Trial, Google's robots placed 1st, 2nd and and 4th.

Wow... Nothing to do with their autonomous cars, which will add tremendous code for aircraft.

To think, google wants autonomous cars so we can spend more time web surfing...    How the economy has changed that advertising can pay for hundreds of millions of lines of new code...

Quoting wowpeter (Reply 57):
No? This is because computer are stupid...

Old computers are stupid. Look at post #55. The computer your discussing wouldn't have had even a hundred thousand lines of code. What we're discussing here is aircraft with a quarter billion lines of code or so.

I worked on newer UAVs with now out of date processors. The amount of code that could be processed today is easily 20X more than what was developed even 3 years ago... Yes, computers are moving forward *that* fast for low power computing.

Quoting wowpeter (Reply 57):
And btw, I fly an Airbus 330... and I can tell you, their software is crap... even the most basic thing... Sigh... only if the average people know this...

Computers were 'crap' in 1990, which is the latest the A330 computer architecture would have been frozen! No wonder you had problems. Seriously, have you noticed how much better computers and software have become in the last 23 years?

The Dec Alpha and PowerPC were introduced in 1992. These were the contenting top architectures for aircraft after the A330/777. In fact, the 777 was bleeding edge using the 486 for a 1994 first flight. The A330, being launched two years earlier, uses something slower (I don't know the exact processor.) Amazingly, the 777 uses the circa 1990 68040 as its backup processor. FWIW, The 486 was new for PCs in 1989. Usually aircraft certified processors trail by 3 years or so, but Motorola used to have chips ready for testing a year before launch, so that is how Boeing safely utilized them for the 777. So pick a circa 1988 processor and de-clock it for the A330. e.g., 386 or 68030.

Let's not compare absolutely out date computers. My 486 with an amazing amount of ram (for the day) struggled to write my masters thesis. What I can do today in a fraction of the time and effort is far more. While doing that work I play music, have a web browser going, text messaging, usually a large background simulation (engineering), dozens of large excel sheets, and a few dozen pdfs open while I'm simultaneously doing a presentation to go with that report. And then sometimes I'll skype on top of that... Now my 140W high clock rate multi core workstation processor won't work in aircraft (they need low power, sub 2 Watt, processors and memory). But the new Octa cores aren't too far off the mark...

Quoting planemaker (Reply 58):
When was the A330 designed? Even the AHMS on the A380 is old in tech terms.

Hey, in 1992 (A330 EIS)I drove a car with a carburetor. Can I blame the difficulty in starting (why the computer didn't even know you had to pump the gas to start...), throttle transients (why you had to anticipate throttle transitions), and the gears required skill to shift gears. Why, it took thought to push in a clutch and select the gear and you could drive so much faster if you pre-selected the gear and anticipated the 'line' for the turn.... And back then you didn't buy electric window lifters unless you were ready to repair them pretty often.

And... I had these things called 'cassette tapes' where I could 'DRIVE' listening to ZZ top man!    Back then, the 'option' for anti-lock breaks was just becoming standard on nicer cars!   

   Sorry, 1992 was a *good* year for me. Thinking back that far brought back a lot of good memories.

But that mac I used only had a 16Mhz processor, but it was a 68030 baby!    And it had a million bits of RAM! (128k) that was upgraded to a whole 4 million bits of ram (512K) by 1992. Anyone remember the first Sim City?   

http://en.wikipedia.org/wiki/Macintosh_IIcx


Lightsaber
IM messages to mods on warnings and bans will be ignored and nasty ones will result in a ban.
 
Pihero
Posts: 4318
Joined: Mon Jan 31, 2005 5:11 am

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 9:04 pm

... and we're still waiting for the intelligent computer   
... while you discuss super processors which won't be used in quite a while in civil aviation (the reasonh one finds 386 and 486 in modern airliners is not a question of cost, but of fiability and heat-robusrtness ).

As I'm quite an open sort of guy, I dived a bit deeper in the research and was impressed with the amount of reseach on robotics ( a great deal of it concerns her favourite subject, the UAVs, and I was even more impressed with the amount of publications they've released into tyhe public domain.
In fact, Ms Cummings work is mostly demonstrates the human / machine interface problems are the way to proceed at this time. The full automaton, she doesn't even seriously envisage.

As driverless cars are one of the so'called proofs of the birth of a new intelligent species by some posters, this is what she and her colleague J. Ryan write :

"Technology Robustness

Because much of the development of driverless cars is proprietary, we do not know their exact capabilities at this time. As such, we cannot make definitive statements about a specific vehicle and can only comment on the limitations of the technology overall and outline specific questions of concern. As best we understand, Google’s autonomous car relies on four major technologies in its autonomous operations: LIDAR (Light Detection and Ranging), a set of onboard cameras, GPS, and stored maps in the vehicle’s onboard computer. The GPS signal tells the car where it is on the stored map and where its final destination is, and from this, the car determines its route. Cameras and LIDAR help the vehicle sense where it is on the road, where other vehicles are, and where to find and follow stop signs and streetlights.

Each of these systems is vulnerable in some form or fashion and it is not clear whether any redundancy exists, or if any one of the four systems fails (maps, GPS, cameras, LIDAR), the car will not be able to function correctly. If the GPS or maps fail, the car does not know where it is on its route and where it should be going. If the LIDAR fails, it may not be able to detect other nearby cars, pedestrians, etc. If the cameras fail, it may not be able to recognize a stop sign or the current color of the traffic light. In addition, it is not clear how much advanced mapping is required by driverless cars and the frequency of map updates that are required to maintain an effective 3D world model by which the onboard computer makes decisions.

The security of GPS has also been questioned (Volpe National Transportation Systems Center 2001, Humphreys, Ledvina et al. 2008). GPS “spoofing” (mimicking the GPS signal to provide false location information) and jamming (forcibly denying GPS signal) attacks have already been observed in US military operations (Franceschi- Bicchiera 2012, Waterman 2012) as well as in civilian applications (Marks 2012). It would not be far-fetched to imagine an individual or group of individuals spoofing GPS signals in major metropolitan areas during rush hour and forcing cars off the road into buildings, off bridges, or otherwise causing damage.

Google’s own researchers admit that inclement weather and construction areas are something they have yet to master (Urmson 2012). Precipitation, fog, and dust are known problems for LIDAR sensors, which can interfere with the image detection capabilities of the camera and can scatter or block the laser beams sent out by the LIDAR. Cameras are also sensitive to such problems. This leaves the vehicle unable to sense the distance to other cars and unable to recognize stop signs, traffic lights, and pedestrians. Urmson also notes that the technology cannot currently handle construction signs, traffic cops (which requires sophisticated gesture recognition that is still an immature technology), and other nonnormal driving conditions. A related question is how well the system can anticipate the actions of other drivers; it is one thing to be able to avoid a car calmly changing lanes, and another entirely to anticipate the actions of a reckless and irrational driver. Given that prior research has shown that people are prone to distraction, any failures or degradations in the technology will significantly increase the likelihood of a serious or fatal accident.
"

( MIT paper ; Humans and Automation Laboratory : "Shared Authority Concerns in Automated Driving Applications")

Is there more to say ?

But stay tuned, there is more on the remote pîlot. I need to summarize without betraying the Lady"s conclusions...

[Edited 2014-01-02 13:08:16]
Contrail designer
 
maxpower1954
Posts: 1067
Joined: Thu Sep 25, 2008 1:14 am

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 10:06 pm

While we breathlessly await Pihero's next installment, here's a story in "The New Yorker" by a cognitive scientist who doesn't believe true A.I. is just around the corner.

Top
 
planemaker
Topic Author
Posts: 5411
Joined: Mon Aug 25, 2003 12:53 pm

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 10:08 pm

Quoting lightsaber (Reply 61):
The implications are huge. Even though I try to follow, the development is happening at an amazing pace with new players surprising everyone.

FWIW, my brother works at a robot vendor and both of us are continually surprised at the rate of progress. His company did a software update that *more than doubled* the robot efficiency.    So much so that their customers couldn't staff up fast enough to do the 'support the robot jobs' to help the robots work at their peak efficiency... Classy problem to have.

The neat thing is that the rate is going to continue.

Quoting FlyHossD (Reply 59):
Nice side-step, I think.

I'm not sidestepping. The priority for the FAA is avoiding UAV - airliner collisions.

Quoting FlyHossD (Reply 59):
One more thing, ALL of my old friends are engineers and apparently good ones, at that (they've had very successful careers). To a man, they all exhibit the "engineer's arrogance" that I'm seeing here. That is, the engineers/designers ALWAYS overlook something - it's just a fact of life.

Honestly, no arrogance. No one is saying that we are going to simply hit the switch and - voila - autonomous airliners. It is a very conservative and logical progression.

Quoting FlyPNS1 (Reply 60):
If it was easily solved, we'd already have a solution. You fail to comprehend some of the legal and regulatory complexities and how slow the FAA moves.

The legal and regulatory "complexities" are straight forward. That the FAA drags its feet is SOP... and that is why congress has and will force the FAA to move faster than it would like to.

Quoting FlyPNS1 (Reply 60):
what you described is really the tip of the iceberg.

It is the UAV industry that says so... and it is not just farmer fields that they are talking about.

Quoting lightsaber (Reply 61):
Hey, in 1992 (A330 EIS)

Ah... but I didn't say EIS... I said designed as that reflects the true tech vintage. So was ~1988 a good year for you?  
Quoting lightsaber (Reply 61):
Sorry, 1992 was a *good* year for me.

It was a *great* year for me. Extended work year in Europe highlighted by attending the '92 Olympics and watching the 1st Dream Team play from court-side. 
Nationalism is an infantile disease. It is the measles of mankind. - A. Einstein
 
User avatar
lightsaber
Moderator
Posts: 18672
Joined: Wed Jan 19, 2005 10:55 pm

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 10:09 pm

Just a thought.

The A330 CPU is most likely a Motorola 68030 or similar. The 68030 has 273 transistors.   

Today's darling cell phone cpu is the ARM A15 with 26M transistors or 100X the processing power per CPU and Planemaker already listed more efficient/higher power alternatives....

The 68030 has a pair of 256 bite cashes. Yes, not kilo or mega bite size, but bites..

http://en.wikipedia.org/wiki/Motorola_68030
IM messages to mods on warnings and bans will be ignored and nasty ones will result in a ban.
 
Pihero
Posts: 4318
Joined: Mon Jan 31, 2005 5:11 am

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 10:45 pm

Quoting lightsaber (Reply 65):
Today's darling cell phone cpu is the ARM A15 with 26M transistors or 100X the processing power per CPU and Planemaker already listed more efficient/higher power alternatives....

...and that will make an idiot computer 10^5 times more idiotic and capable of being 100 times faster in idiocy... some performance !   
Still on power and not intelligence, hey ?

[Edited 2014-01-02 14:49:56]
Contrail designer
 
maxpower1954
Posts: 1067
Joined: Thu Sep 25, 2008 1:14 am

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 10:56 pm

Couldn't link this before - not all scientists think A.I. is right around the corner.

Top
 
maxpower1954
Posts: 1067
Joined: Thu Sep 25, 2008 1:14 am

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 11:02 pm

Still can't link it! Very interesting article by a cognitive scientist on why he thinks true A.I. is years away.

Google search - Gary Marcus Hyping Artificial Intelligence
 
planemaker
Topic Author
Posts: 5411
Joined: Mon Aug 25, 2003 12:53 pm

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 11:53 pm

Quoting maxpower1954 (Reply 67):

Couldn't link this before - not all scientists think A.I. is right around the corner.

No one has said that all scientist believe that AI is right around the corner (much less a psychology professor like Marcus). The consensus is that it will be achieved by ~2050 while others, like Kurzweil at Google, believe that it can happen as soon as 2029. Where we are located at the bend of the knee is what is being debated. But at the end of the day, the timing over full-AI is a canard... it isn't required.
Nationalism is an infantile disease. It is the measles of mankind. - A. Einstein
 
prebennorholm
Posts: 7020
Joined: Tue Mar 21, 2000 6:25 am

RE: Flight Global: When Will We Drop The Pilot?

Thu Jan 02, 2014 11:56 pm

Pilotless airliners will happen no earlier than the day when ground based voters have convinced the ground based politicians that they can easily vote in favor of them, and still be reelected. That means never.

In the meantime we will over the next few centuries see UAVs performing tasks which we today can hardly dream about. And all of them will have a ground based controller with his right hand index finger on a small red button to push when they head south.
Always keep your number of landings equal to your number of take-offs
 
planemaker
Topic Author
Posts: 5411
Joined: Mon Aug 25, 2003 12:53 pm

RE: Flight Global: When Will We Drop The Pilot?

Fri Jan 03, 2014 12:12 am

Quoting prebennorholm (Reply 70):
Pilotless airliners will happen no earlier than the day when ground based voters have convinced the ground based politicians that they can easily vote in favor of them, and still be reelected. That means never.

In the land of the free and home of the brave   it will be the consumer that will eventually chose. And considering how ULCC's have done, a sizable portion of the population vote with their wallets. As Prof Missy Cummings reports, when she asks her class who would fly in a pilot-less aircraft all the students put up their hands. I guess it is a generational thing.  
Nationalism is an infantile disease. It is the measles of mankind. - A. Einstein
 
UA772IAD
Posts: 1321
Joined: Sat Jul 10, 2004 7:43 am

RE: Flight Global: When Will We Drop The Pilot?

Fri Jan 03, 2014 1:26 am

Quoting lightsaber (Reply 31):
Trains are the perfect example! Look how much more efficient they are. Once upon a time one would have an engineer, assistant engineer, a stoker, and a fireman (or two) then at there would be 3 to 5 breakmen. Eventually it went to two engineers and one breakman in the caboose. Now it is one engineer (a 2nd if a trainee) for the train and many subways have remote control (no human).

Sorry but this is not accurate. Heavy rail (Class I) railroads operate with at least two crew: the engineer and the conductor. The conductor is responsible for conveying and ensuring that all radio communication, track warrants and special operating circumstances are adhered to during the operation of the train. They are also a second set of eyes, ensuring signals, speed restrictions and crossings are safely negotiated.

Automated Train Control, or Positive Train Control as established in mass transit isn't remote control either. The subway cars have sophisticated computers onboard which can operate the train as programmed- however, there is still an operator in the cab. The operator ensures that the ATC/PTC does what its supposed to, but there are many instances where ATC/PTC cannot be used- such as in a work zone, or where single-tracking may occur, and thus the operator needs to take physical control and operate the train.

Quoting planemaker (Reply 3):
As the article stated, there will be a generational shift to acceptance. How it unfolds exactly no one can predict other than it will happen

If no one can predict how it will happen, how can anyone definitively say that it will happen? Humans aren't so one dimensional, we don't just accept advancing technology as a sign of the times and go with it unquestioned. The history of technology is filled with examples.
 
planemaker
Topic Author
Posts: 5411
Joined: Mon Aug 25, 2003 12:53 pm

RE: Flight Global: When Will We Drop The Pilot?

Fri Jan 03, 2014 1:40 am

Quoting UA772IAD (Reply 72):
If no one can predict how it will happen

I used the word exactly... and why no one can predict "how it unfolds exactly" is that various approaches are being considered. In other words, there are options and it is too early to specify which path will be chosen.
Nationalism is an infantile disease. It is the measles of mankind. - A. Einstein
 
User avatar
lightsaber
Moderator
Posts: 18672
Joined: Wed Jan 19, 2005 10:55 pm

RE: Flight Global: When Will We Drop The Pilot?

Fri Jan 03, 2014 2:42 am

I'll post the FAA roadmap for unmanned vehicles in the NAS:
http://www.faa.gov/about/initiatives/uas/media/uas_roadmap_2013.pdf

Look at Appendix C and all the projects that where held back by the FAA until the release of that document.   
Three UAS manufacturers have already applied for type certification and two of these
applications were released from delayed sequencing to proceed with restricted category airworthiness
certification.


That is three applications for NSA approved UAS vehicles.   

I do think certifying a UAV for operation sans visual observer in 2016 to 2020 is very pessimistic of the FAA (see section C.4 of the link).

NASA has more press on their UAV airspace integration contest (already posted):
http://www.executivegov.com/2014/01/...-uav-airspace-integration-contest/


Part of my opinion is based on how many serial flights a certain 'piloted UAV' has performed where the pilot has done nothing but be baggage. Where we are now is *far* more advanced than 2011.

Same is true with many robots, but I know things through friends who hit new efficiency benchmarks and not via links...

Note: this thread has pointed out quite a bit of development I wasn't aware of.   

Quoting Pihero (Reply 66):
Still on power

Do you code? I get the idea you do not understand how running more subroutines in parallel enhances predictive ability. This is standard stuff when testing new aircraft. If you missed something... add a subroutine to make sure you don't again. Insure redundancy and error checking.

Do you know how UAV avionics goes through its 'voting process' to decide what to do? When you add more complexity, right now we have to add dedicated processors to handle new capabilities. e.g., complex flying is already being performed by UAVs:
http://www.youtube.com/watch?v=FMbwifwj5pg
http://www.youtube.com/watch?v=_zC8diFd5O8

Now the videos released are very censored. In the first video, realize those are two aircraft flying *incredibly* close to their stall limits for that demonstration. (Too close for a human pilot.) They have little room to accelerate or decelerate relative to each other due to the extreme altitude with the global Hawk refueling. But look how they maneuver!   

The second video show UCAS getting its initial carrier certification. What it did then wasn't pushing the envelope. That plane will be able to operate off a carrier in weather you couldn't fly a human piloted plane. And UCAS is absolutely primitive in its avionics vs. UCLASS (its replacement, UCAS is only a demonstrator while UCLASS is still being planned). When I get excited about improved processing power... its because I was doing the system center lab testing and I knew what wasn't being put in due to thermal loads. That is the hard part of developing new aircraft. Will problems be found in the field? Maybe. So much is found in the labs today.

What is shown in those videos was considered beyond computer capability in 2011... What will be done in 2014 will make those videos look like child's play. Heck, the best stuff those aircraft did in 2012 was censored from the videos (gave away too much of aircraft capability).

Quoting planemaker (Reply 69):
No one has said that all scientist believe that AI is right around the corner (much less a psychology professor like Marcus).

Exactly. Most of this is augmentation, not replacement. Even UCLASS won't replace human piloted navy planes. But... the F-35 is almost certainly the last piloted Navy fighter (but it will be around for decades).

Quoting UA772IAD (Reply 72):
Sorry but this is not accurate. Heavy rail (Class I) railroads operate with at least two crew

Interesting, I found many links on solo engineers. However, all were written with a bias to creating more jobs on the issues rather than it being practice. Apparently the pro-solo engineer crowd is just quietly implementing the work.
http://www.ble-t.org/pr/newsletter/1299newsletter/story6.html
Union Pacific management has imposed a policy that allows engineers only two familiarization trips over the territory between Portland and Hinkle, Ore., then forces them to operate their train solo.

I posted my prior links. Your railroad might still require a conductor, but if there aren't solo engineers, why did I just find so many links on the practice? And yes, there is an engineer if rail projects aren't in condition for autonomous operation.    We're getting there. I've ridden rail where the operator was remote. That's how it should be done. Aircraft will not only have autonomy, but I strongly believe remote copilots.

Its ok to have an opinion, but post links to prove your case. Since I find solo engineers a labor issue... that is being implemented, I find your assertion incorrect.

Quoting Pihero (Reply 62):
Each of these systems is vulnerable in some form or fashion and it is not clear whether any redundancy exists

Your kidding with regard to aviation right? Anyone who has taken even a basic avionics course knows that multiple redundancy is required. Often triple or quad redundancy in software. Most of the aircraft grandfathered in to keep flying would never pass today's system center lab testing. The software testing is now that rigorous. Heck, we have fun testing scenarios where a piloted plane would always die...

But I'm not talking going to be able to explain system center software testing in a thread.

Quoting UA772IAD (Reply 72):
Humans aren't so one dimensional, we don't just accept advancing technology as a sign of the times and go with it unquestioned. The history of technology is filled with examples.

What's your point? Unions fought 'lights out' machining for decades... until it was common. The most efficient producer will be the one still in business.

Quoting planemaker (Reply 64):
No one is saying that we are going to simply hit the switch and - voila - autonomous airliners. It is a very conservative and logical progression.

The motto is crawl, walk, run with UAVs. Since that concept was adopted, progress has accelerated as the software building blocks were laid.

Quoting planemaker (Reply 64):
That the FAA drags its feet is SOP... and that is why congress has and will force the FAA to move faster than it would like to.

If congress hadn't pushed aside the objections of the 'old guard,' the USA could seriously have fallen behind several other nations in UAV technology.

The FAA is getting hell with how slow they moved in 2013. There is discussion of pulling some of their funding and moving it (where? I don't know) due to their foot dragging on UAVs. Even talk of removing some airspace from the NAS to accelerate development. There is being conservative and then there is being obstructive.

Six new *commercial* UAS flight test centers will make quite a difference.   

Quoting planemaker (Reply 64):
So was ~1988 a good year for you?

Yes, but 1987, 1992, and 1999 were my best years.   

The good old 68030 baby!    (Is that the A330 processor? It would have been logical for that era.)

Lightsaber
IM messages to mods on warnings and bans will be ignored and nasty ones will result in a ban.
 
planemaker
Topic Author
Posts: 5411
Joined: Mon Aug 25, 2003 12:53 pm

RE: Flight Global: When Will We Drop The Pilot?

Fri Jan 03, 2014 4:41 am

Quoting lightsaber (Reply 74):
The second video show UCAS getting its initial carrier certification. What it did then wasn't pushing the envelope. That plane will be able to operate off a carrier in weather you couldn't fly a human piloted plane. And UCAS is absolutely primitive in its avionics vs. UCLASS (its replacement, UCAS is only a demonstrator while UCLASS is still being planned). When I get excited about improved processing power... its because I was doing the system center lab testing and I knew what wasn't being put in due to thermal loads. That is the hard part of developing new aircraft. Will problems be found in the field? Maybe. So much is found in the labs today.

What is shown in those videos was considered beyond computer capability in 2011... What will be done in 2014 will make those videos look like child's play. Heck, the best stuff those aircraft did in 2012 was censored from the videos (gave away too much of aircraft capability).

Nice videos... and nice work you were doing!!  
Quoting lightsaber (Reply 74):
The good old 68030 baby!    (Is that the A330 processor? It would have been logical for that era.)

Oh, you are good... it is the 68010... and the 80186 (two different ones for redundancy).
Nationalism is an infantile disease. It is the measles of mankind. - A. Einstein
 
rwessel
Posts: 2448
Joined: Tue Jan 16, 2007 3:47 pm

RE: Flight Global: When Will We Drop The Pilot?

Fri Jan 03, 2014 1:16 pm

Quoting wowpeter (Reply 57):
For your info, if Swireair were to divert and land with an overweight landing, they would have live, but they elect to dump fuel, wasting precious time... Pilot are now taught not to do that... Just go overweight landing in cargo fire... you only have 15 mins or else you will need to put it in the water or on the ground.

People have run the numbers on that several ways, but it's pretty improbable. To get on the ground, they would have had to divert at the very first sign of smoke, and then done busted all the speed limits and whatnot on they way to Halifax. That's the TSB's official conclusion as well - a successful landing would not have been possible, there simply was not enough time before the aircraft became uncontrollable.
 
User avatar
lightsaber
Moderator
Posts: 18672
Joined: Wed Jan 19, 2005 10:55 pm

RE: Flight Global: When Will We Drop The Pilot?

Fri Jan 03, 2014 2:22 pm

Quoting planemaker (Reply 64):
The neat thing is that the rate is going to continue.

In my opinion, the fastest rate of acceleration will be the next 3 years then it will be like computers in the 1990s... steady progress that dramatically improved their capabilities.

My youngest child loves robots and we're encouraging that. Her favorite Christmas toy was a simple robot, but its already obvious that for her birthday she'll need something very sophisticated. I just found an age appropriate robot getting rave reviews on Amazon:
http://www.amazon.com/Kid-Galaxy-104...d=1388758602&sr=8-3&keywords=robot

Quoting planemaker (Reply 75):
Quoting lightsaber (Reply 74):
The good old 68030 baby!    (Is that the A330 processor? It would have been logical for that era.)

Oh, you are good... it is the 68010... and the 80186 (two different ones for redundancy).

Really?!? Wow. Amazingly efficient coding then. The French are most excellent at software logic IMHO. That processor, IIRC, has less than 200k transistors.    My kids toys have multitudes more processing power.

Lightsaber
IM messages to mods on warnings and bans will be ignored and nasty ones will result in a ban.
 
UA772IAD
Posts: 1321
Joined: Sat Jul 10, 2004 7:43 am

RE: Flight Global: When Will We Drop The Pilot?

Fri Jan 03, 2014 4:16 pm

Quoting lightsaber (Reply 74):
I posted my prior links. Your railroad might still require a conductor, but if there aren't solo engineers, why did I just find so many links on the practice? And yes, there is an engineer if rail projects aren't in condition for autonomous operation.    We're getting there. I've ridden rail where the operator was remote. That's how it should be done. Aircraft will not only have autonomy, but I strongly believe remote copilots.

Because you took what I wrote, disregarded half of what I said and are arguing a moot point. I never said there weren't single engineers operating trains, I said the engineer isn't the only one up in the cab.

The conductor does not operate the controls of the train. The engineer does. The conductor is responsible for the train consist itself. You will not find a Union Pacific, CSXT, BNSF, Norfolk Southern or Amtrak locomotive with just one body in the cab. There will always be a conductor and an engineer.

"My Railroads" all employ conductors:
http://www.csx.com/index.cfm/working.../transportation/freight-conductor/
https://up.jobs/job/opening/Train%20Crew/Saint%20Paul/MN/071603?jsl=2974517
http://www.bnsf.com/careers/explore-...ades/transportation/conductor.html
http://www.nscorp.com/content/nscorp...career-paths/hourly-positions.html

Quoting lightsaber (Reply 74):
Its ok to have an opinion, but post links to prove your case. Since I find solo engineers a labor issue... that is being implemented, I find your assertion incorrect.

See above. And thanks for allowing me to have an "opinion"

Quoting lightsaber (Reply 74):
What's your point? Unions fought 'lights out' machining for decades... until it was common. The most efficient producer will be the one still in business.

My point is, this is a reductionist argument that can't be proven. By the way, I'm not making this a labor issue. Why are you?
 
planemaker
Topic Author
Posts: 5411
Joined: Mon Aug 25, 2003 12:53 pm

RE: Flight Global: When Will We Drop The Pilot?

Fri Jan 03, 2014 8:14 pm

Quoting lightsaber (Reply 65):
Today's darling cell phone cpu is the ARM A15 with 26M transistors or 100X the processing power per CPU

People are so fickle... they are already looking at the new hottie... the 64-bit A50 series.  

I don't think that people appreciate the price/performance of their cell phones. It is objectively truly amazing. Since we were joking about the original processor on the A330, it is now hard to believe that GPS wasn't available at EIS... kinda take GPS so for granted with it in our cell phones.

And speaking of cell phone utility, here is a video showing how the optical component in a cell phone (3:50 min mark) is used in a partially classified surveillance imaging platform.  
Quoting lightsaber (Reply 74):
The FAA is getting hell with how slow they moved in 2013. There is discussion of pulling some of their funding and moving it (where? I don't know) due to their foot dragging on UAVs. Even talk of removing some airspace from the NAS to accelerate development. There is being conservative and then there is being obstructive.

One UAV manufacturer in Washington has been going up to Canada for testing as it takes only 14 days to get a test permit.

Quoting lightsaber (Reply 77):
In my opinion, the fastest rate of acceleration will be the next 3 years then it will be like computers in the 1990s... steady progress that dramatically improved their capabilities.

That might very well be the case. However, there are several factors (wild cards?) that could fairly easily enable the "acceleration" to really continue.

[quote=lightsaber,reply=77]IMHO. That processor, IIRC, has less than 200k transistors.

I've seen it quoted a few times that a $2 musical greeting card has more computing power than the Apollo 11 Lunar Lander... and we toss the card in the garbage. 
Nationalism is an infantile disease. It is the measles of mankind. - A. Einstein
 
KD5MDK
Posts: 816
Joined: Wed Mar 13, 2013 4:05 am

RE: Flight Global: When Will We Drop The Pilot?

Fri Jan 03, 2014 8:34 pm

 
User avatar
hOMSaR
Posts: 2254
Joined: Tue Jan 19, 2010 4:47 am

RE: Flight Global: When Will We Drop The Pilot?

Sat Jan 04, 2014 1:13 am

Quoting UA772IAD (Reply 78):
You will not find a Union Pacific, CSXT, BNSF, Norfolk Southern or Amtrak locomotive with just one body in the cab.

Plenty of Amtrak trains run with just one man in the cab. The conductor still exists, but he's in the back of the train dealing with the passengers.
The plural of Airbus is Airbuses. Airbii is not a word.
There is no 787-800, nor 787-900 or 747-800. It's 787-8, 787-9, and 747-8.
A321neoLR is also unnecessary. It's simply A321LR.
Airplanes don't have isles, they have aisles.
 
User avatar
lightsaber
Moderator
Posts: 18672
Joined: Wed Jan 19, 2005 10:55 pm

RE: Flight Global: When Will We Drop The Pilot?

Sat Jan 04, 2014 4:08 am

Quoting homsar (Reply 81):
Plenty of Amtrak trains run with just one man in the cab.

Thank you, very interesting. But in my opinion, the conductor on the Amtrak train has a role with passengers analogous to a pursuer on a cruise ship; it goes beyond the definition of what is required to operate the vehicle and is more part of the 'expected passenger experience' that includes a significant safety role analogous to flight attendants.

Quoting planemaker (Reply 79):
People are so fickle... they are already looking at the new hottie... the 64-bit A50 series.

My cell phone is only a few months old and I'm already looking forward to the not only the A50 series, but what AMD, Qualcomm, and Samsung will do with their own 64 bit designs.   

Quoting planemaker (Reply 79):
I don't think that people appreciate the price/performance of their cell phones. It is objectively truly amazing.

I'm floored at the progress. Some of the cell phone technology blew past aerospace standards so quickly that established R&D labs we left explaining how 'cheap' commercial technology passed them by so quickly. It is amazing what volume can do for R&D progress.

But I feel old!    I see the interns using smartphones for shortcuts that I just do not consider. While I'm learning the tricks, it amazes me what latent capabilities lie within the latest generation of smart phones. I still haven't picked up the habit of *not* taking notes but rather just taking pictures of everything... e.g., today we knew why a part malfunctioned because the intern photographed assembly utilizing his smartphone in amazing detail. Enough detail that we *knew* without opening the part that the output signal calibration was mis-set.

I'm still of the mind set of a few years ago that photos take time and effort...    Not of the mindset of taking 30 extra photos and putting them in a folder 'just in case.' But rather of the mindset that each photo has a documentation purpose instead of 'take the photo, it might have useful information.'

And then GPS, tapping to trade information, and having the right bevy of apps...

Who would have thought that having a bunch of small programs in your hand could be so useful? Touch interfaces, apps, cameras, GPS, bluetooth, near-field communication, and fast data systems for so little is really changing everything.

4 years ago when I told all my aerospace buddies that we would be riding on the tails of cell-phone innovation created chuckles. Now it creates stares of 'how do we keep up.'

If anyone finds a reason to put a cheap IR camera in a smartphone... that would lay the groundwork to change a dozen industries...

Quoting planemaker (Reply 79):
One UAV manufacturer in Washington has been going up to Canada for testing as it takes only 14 days to get a test permit.

I wish I could say I'm surprised... oh, I didn't know about it, but it seems only logical with the FAA foot dragging.

Quoting UA772IAD (Reply 78):
My point is, this is a reductionist argument that can't be proven. By the way, I'm not making this a labor issue. Why are you?

My point is that automation has reduced labor needs in other transportation. As to labor issue, I just noted which web sites had the most information. For some reason those implementing the solutions are being very quiet. Excluding the Dubai metro that is:
http://en.wikipedia.org/wiki/Dubai_Metro

Here is the list of automated trains I posted before:
http://en.wikipedia.org/wiki/List_of_driverless_trains
n an Unattended Train Operation (UTO) system like the Copenhagen Metro, the trains run automatically at all times, handle door closing, obstacle detection and emergency situations, without any regulatory requirement of staff present in the trains.

Quoting UA772IAD (Reply 78):
See above. And thanks for allowing me to have an "opinion"

I encourage discussion here on a.net. I like opinions different than my own. Yet you stated as a fact (my statement was not accurate), not as an opinion. I provide links to prove my point that there are "Grade of automation 4" trains which indeed there are. I'm sure many railroads employ more people just as some airlines purchased 3 pilot cockpits after the industry started the transition to 2 pilot cockpits. The proof that autonomous operation works is a few examples that operate. If someone wants to buy an autonomous system and operate it at a reduced level of automation?    But we have Grade 2, 3, and 4 systems in service. What I propose is we'll have Grade 2 and 3 automation for transport aircraft.

The neat thing about automation is once it is installed, software upgrades will allow improvement in efficiency, safety, timeliness, information transfer, durability, and diagnosis. I know many on here like to point out the 787 faults, but once it is debugged, it will be the lowest cost widebody out their to maintain for flight in part as the components will let the right people know a few hundred flights beforehand that maintenance needs to be scheduled. I've seen the progress on other aircraft. At first the computers are frustrating... but once debugged, everything works and the computers are letting you know about potential areas of improvement.

The same is true of autonomous flight augmenting a solo pilot. The computers can let the pilot know about needed corrections long before a human could react. I've seen how far out the cameras/computers can detect possible collisions and correct far outside human reaction range without all but the experts knowing the aircraft corrected. I don't know if I should laugh or cry knowing this safety capability came from modified targeting algorithms...

Quoting planemaker (Reply 79):
I've seen it quoted a few times that a $2 musical greeting card has more computing power than the Apollo 11 Lunar Lander... and we toss the card in the garbage.

I just threw away a cell phone that had more computing power, by about a factor of 30, than a $3,500 computer I once splurged on because I needed a 'true engineering workstation'... My new cell phone puts to shame a $60k Sun workstation I once had at a prior employer.

I'd buy one of the 'internet fridges' if I didn't know in two years it would be totally obsolete and the new one at half the price...

Lightsaber
IM messages to mods on warnings and bans will be ignored and nasty ones will result in a ban.
 
rwessel
Posts: 2448
Joined: Tue Jan 16, 2007 3:47 pm

RE: Flight Global: When Will We Drop The Pilot?

Sat Jan 04, 2014 7:44 am

Quoting lightsaber (Reply 82):
If anyone finds a reason to put a cheap IR camera in a smartphone... that would lay the groundwork to change a dozen industries...

A certain level of IR sensitivity is actually natural for many of the sensors used in digital cameras. That's actively filtered out because more than a bit of clothing is semi-transparent in near-IR, which leads to something of a PR disaster. So there may be some resistance to actively integrating good IR in general devices.

OTOH, fairly inexpensive near-IR sensors do exist, although they haven't seen near the effort that visible wavelength devices have seen to reduce size, power consumption and cost (while simultaneously improving resolution and speed), simply because the volume isn't there.

The further from the visible spectrum you get the more expensive things become, although again, a fair bit of that is volume related.

But OTTH, a "night vision" mode on your cell phone might well be useful at times - certainly such a thing on your (say) car would be useful.

But in general if you get the infinite volume end of the industry interested, can get huge improvement in a technology. Although they can be fickle. Witness the decline of serial ports, which used to be a super cheap way to interface a PC to hardware - serial ports have largely disappeared from laptops, and more than a few desktops don't have them either. Adding them is trickier than you might think, since the USB/serial adapters almost all have a number of issues that makes them difficult to use for anything other than modem-like devices (but if you can add a PCIe or ExpressCard device, those usually work just fine).

Still, a decade ago who'd have thought we'd have (near) inertial nav systems on our cell phones.
 
tim73
Posts: 93
Joined: Sun Nov 17, 2013 9:03 pm

RE: Flight Global: When Will We Drop The Pilot?

Sat Jan 04, 2014 11:49 am

Quoting lightsaber (Reply 82):
Some of the cell phone technology blew past aerospace standards so quickly that established R&D labs we left explaining how 'cheap' commercial technology passed them by so quickly. It is amazing what volume can do for R&D progress.

I am pretty sure, the cell network&phone R&D has been the number one technology when it comes to money spending, starting from around 1992-1993. It required so many breakthrough innovations via expensive research projects. No wonder the fruits of those projects are now spreading to many other areas too.

It is kind of funny, any DYI amateur could build a better small surveillance drone with things like Nokia 41 megapixel monster Pureview 808 optics  

It would be interesting to see, if somebody really used the latest technology to build the most advanced, pilotless passenger plane ever, without going through the tedious certification processes   Maybe those processes are a bit outdated?
 
peterjohns
Posts: 176
Joined: Fri Jan 02, 2009 11:49 am

RE: Flight Global: When Will We Drop The Pilot?

Sat Jan 04, 2014 4:01 pm

Interesting Topic.
Also interesting why nobody comes up with why it WON´T work in the foreseeable future.
It may very well be technically probable, but would it save money!!??

As an ATCO myself I can see many reasons why decision authorized persons (Pilots?) must be on board an a/c.
- medical emergencies, - diversions , resetting the computers in flight (ever flown an Airbus? - circuit breakers)...
weather avoidance, level change due to turbulance, VFR avoidance when flying into according airspace (law)
 
User avatar
SuseJ772
Posts: 935
Joined: Wed Aug 10, 2005 11:13 am

RE: Flight Global: When Will We Drop The Pilot?

Sat Jan 04, 2014 5:36 pm

I'd like to thank Lightsaber and Planemaker for all the information in this. As a CTO and a private pilot, I found your insights very interesting. I like to think one of my skills in life is sitting in the balance of technology innovation, cost effectiveness with ROI and efficiencies. Needless to say, this topic interests me greatly and I have learned a lot from you today.

Quoting EA CO AS (Reply 9):
You'd sooner see the safety component of flight attendants done away with and them replaced with onboard vending machines than you'll see a commercial pilot removed from the flight deck.

I would think it would be easier to get rid of the pilot than it would be the flight attendant as all flight parameters can be programmed for. Getting humans off of a plane in an emergency cannot be.

Quoting rwessel (Reply 14):
OTOH, the guy on the ground isn't facing the distraction of his own imminent demise...

He is if he loses his job because of it.

Quoting Rara (Reply 18):
AF447 is probably the worst argument in favour of human pilots you could come up with.

Agree. I have to believe if people weren't in such shock, they would have noticed the extreme descent and slow airspeed.

Quoting homsar (Reply 20):
There are also onboard fires that kill pilots long before the plane can make it back to an airfield. If, by some chance, the plane is still flyable, the computer could get the plane on the ground safely (maybe important when flying over populated areas with potential ground casualties) whereas the pilots, being incapacitated, would not.

Exactly. Like UPS 6.

Quoting Indy (Reply 21):
If a drone pilot crashes a plane he just gets a new plane.

I think that will be why the law should require that if this becomes a reality, the pilot will lose his job.

Quoting JBirdAV8r (Reply 24):
1) Cost of certification ($$$$$).

This is myopic thinking. First as some have stated, most of this is already being done and will be done before it gets to commercial aviation. The second thing is that payroll is the second LARGEST expense behind fuel. It's why most pilots are pilots and most management is management. Pilots usually have no concept of ROI of investment. The idea is if it take 5 billion to certify, but that you can cut your second largest expense by 50%, than it is worth it. If I told an airline that there was technology out there that they had to spend 5 billion on, but it would remove the need for fuel, you bet your hat they would go through whatever is necessary. The same thing with "salaries" - which convienently, most employees don't recognize (and pilots are employees). The cost is expensive, but so are the savings.

Quoting JBirdAV8r (Reply 24):
2) Cost of overhauling infrastructure/designing and certifying new aircraft/bringing these fleets online ($$$$$). Retrofits with the technology to even make some existing airplanes single-pilot capable would run into the millions of dollars per existing airframe, making this cost-prohibitive. That leaves you with waiting for single-pilot or unpiloted airliners to be developed, certified and brought online. With this not even on the radar yet, this is probably at least 40-50 years away.
Same as above.

Quoting Pihero (Reply 36):
That, I’m afraid denotes your total misunderstanding of airline operations. Dispatchers routinely devise our paths based on their knowledge of the traffic and the weather. There is no JFK-SFO route and coded as such.
Pihero, I usually respect you on this board more than just about anyone. But in this thread it feels like you are grasping at keeping your job and sounds desperate. This quote is a perfect example: you are right, there is no "route" from JFK-SFO, but you should know that loading the route and/or changing the route, is not particularly complicated. You do it yourself currently. The only difference is that in the current reality, a ATC person tells you to change the route, and you change the parameters in the computer. If the ATC person (or someone else responsible in the future OR computers just automatically know to change it), can just change it in the computer directly, then the pilot is still not necessary - which is what we are discussing here.

Quoting Pihero (Reply 36):
there is no autopilot than can cope with a simple bounced landing.
It doesn't mean it won't be able to cope with it in the future.

Quoting Nav20 (Reply 45):
Every airport in the world - including alternates along all the routes - would have to be equipped with full ILS etc.?
Not if ILS is not what is used. ILS is a pretty antiquated technology. Yes it has worked well, but it does not mean it is the only option that will be certified in the future. You're still looking in the rearview mirror if you think ILS is the answer here.

Quoting planemaker (Reply 58):
Some people might know that Amazon is the largest online retailer but are completely oblivious to the fact that they are one of the biggest web services provider with clients including the CIA and NASA and that they own a robotic company.
Amazon AWS. It's the other tab open on my browser right now   

Quoting planemaker (Reply 58):
and of course, Motorola.
Which of course is now owned by Google.
Currently at PIE, requesting FWA >> >>
 
Pihero
Posts: 4318
Joined: Mon Jan 31, 2005 5:11 am

RE: Flight Global: When Will We Drop The Pilot?

Sat Jan 04, 2014 9:04 pm

Back to the original subject, shall we ?
First of all, there is the fact that drones heve been with us for a good part of sixty years : reformed fighters or trainers (there were a loit of T-33s, but also, F-86s, 100s...etc...) used as targets..
That means that with the addition of a very limited number of electronics, any modern airliner could be either remotely piloted or fully automated : flying an unrestricted preprogrammed flight path has been after all in the capability of any UAV or any cruise missile in quite a few military inventories.
There are, though, a few problems to overcome before, as the OP so elegantly states, "we drop the pilot" :

- One is the fact that I have never flown a sector without an ATC amendment, i.e ouside the flight plan,
- Two is that I've never been alone in any given airspace
- Three is that I've had to cope with quite a few abnormalities ( even some emergencies) due to system failures, passenger related... which led to some drastic desisions on alternative flight management strategies.
- Four is not an easily dismissed matter about the amount and accuracy of the information I'd need to base a strategy / decision on . I posit that an automaton would require similar info to perform correctly.
- Five, the levels of redundancy have to be completely rethought :

At this time, we define three status levels of redundancy :

- Fail-passive is the ability of the system to detect a fault, provide a warning and automatically disengage the A/P in a manner to limit any acceleration forces or flight path debviations to an insi_gnificant level.
- Fail-soft is the ability to withstand a failure without endangering safety and without producing excessive déviations in the flight path, if reasonable pilot attention is provided
- Fail-operational, defined as the ability of the system to revert to a fail-passive status in the event of a fault, with approprioate warning and with no degradation in A/P performance.

From the above définitions we could - if we are honest - derive at least two remarks :

1/- every automated approach-and-landing will be done in Cat IIIc conditions, whatever the weather - we know what a wimp George is in windy landings - or the visibility ( as we have to insure asurvivable landing with a *blind* automaton at the controls.
That means, for safety sake, that no degradation from Fail-operational will be allowed... resulting in at least another layer of higher level of redundancy ( at what cost ?). And, no the answer is not relying on some hot-shot remote operator for reasons of com bandwidth, lag and secure encryption : the affected aircraft would be spread all over the runway before (s)he could intervene.
That also means that the MEL (minimum Equipment list ) will be a thing of the past as no deferred defects will be acceptable at the launch of the flight (at what cost ?).

2/- AF447 is an example of a degradation from fail-operational to fail-passive... or is it ?...24 simultaneous broadly unrelated failure messages ?.. what was the clear warning of the status of the system ?.. and what was the status of the Flight Director ?.. Wasn't it a major failure that did not remove the FD from giving totally erroneous commands ?..
We enter here the realm of what is forecast by engineers against what could in fact happen... Unforeseen circumstances, therefore not envisaged, hence not programed ( or is the word *coded* ? )

The next question is : " What makes humans and computers different ?"
Fitts lists the différences at task allocation :

Humans are better at:

Perceiving patterns
Improvising and using flexible procedures
Recalling relevant facts at the appropriate time
Reasoning inductively
Exercising judgment

Computers are better at:

Responding quickly to control tasks
Repetitive and routine tasks
Reasoning deductively
Handling simultaneous complex tasks
Fast and accurate computation

It shows that the two are inherently different... It also shows the remaining distance computers need to complete to achieve some degree of *intelligence.
To say that will be done in the next fifteen years (time limit of the Kurzweil - Kapor wager which defines a very simplistic capacity of human relationships as *intelligence* ) is at the very last a very optimistic view of the computer capabilities, or worse : pure intellectual dishonesty.

If I may, I'd cite once again Ms Cummings and her colleagues at M.I.T :

"…in contrast to automation brittleness, humans’ strengths in planning environments are their abilities to improvise, learn, and reason inductively, which are precisely the skills required to adapt to unexpected circumstances. This type of problem solving is called knowledge‐based reasoning ,
during which humans make decisions under novel and uncertain situations, which are attributes inherent to supervisory control scenarios. In terms of managing the large data streams that will be
generated in smart environments, automation will be critical in handling the bulk of problem solving and system management. However, just as evidenced by the 2003 Northeast blackout, even highly automated systems can be presented with a set of dynamic and unexpected variables states never envisioned in advance by designers, which can ultimately lead to catastrophe.
"
... and :
"Contrary to many assumptions, the insertion of more automation, both in terms of distributed sensors and algorithms to post‐process data, will not necessary reduce workload, nor necessarily improve system performance. Several studies reviewed by Parasuraman, et al. suggest that automation in aircraft cockpits has been known to actually increase workload, rather than decrease workload as intended.
Interestingly, there is little evidence supporting the idea that operators will delegate tasks to automation when workload is high.
"

Paying attention to the man behind the curtain

Can't say it better .

And finally, the danger lurking behind the remote pilot :

"While automation has greatly reduced operator workload and generally enhanced safety in supervisory control settings, humans can fall prey to “the ironies and paradoxes of automation”.
Increased automation can lower an operator’s task load to the point where vigilance is negatively affected and boredom possibly results. Unfortunately, as increased automation shifts controllers into system management positions, monotony, loss of vigilance, and boredom are likely to proliferate

“Highly skilled, highly trained people can only eat so many peanut M&Ms or Doritos or whatnot…There's the 10 percent when it goes hot, when you need to shoot to take out a high-value target. And there's the 90 percent of the time that's sheer boredom—12 hours sitting on a house trying to stay awake until someone walks out."
In a recent Predator operations study, 92% of pilots reported‘moderate’ to ‘total’ boredom

Ultimately, we must accept that highly automated environments will likely be considered boring environments that can and will lead to distraction. Thus the question should not necessarily be how to stop distraction but how to manage it through either personnel selection considerations as, or through more active interventions such as designing alerting systems to promote optimal attention management strategies.

Another possible technology solution would be use of some psychophysiological alerting system such as one based on galvanic skin conductance to warn either individual operators or supervisors that a low vigilance state has been achieved…
"
( in other words," We'd put some electrical wire up your butt if you're not paying attention, stupid ! )...

But I seee we're still talking smart phones and smart processors...  

[Edited 2014-01-04 13:11:33]

[Edited 2014-01-04 13:12:38]

[Edited 2014-01-04 13:16:13]
Contrail designer
 
planemaker
Topic Author
Posts: 5411
Joined: Mon Aug 25, 2003 12:53 pm

RE: Flight Global: When Will We Drop The Pilot?

Sat Jan 04, 2014 9:18 pm

Quoting lightsaber (Reply 82):
Thank you, very interesting. But in my opinion, the conductor on the Amtrak train has a role with passengers analogous to a pursuer on a cruise ship;

From one of the Flight Global articles: The aircraft commander will be the Purser – the senior cabin crew member – and the pilot back at base will be the driver.

Quoting lightsaber (Reply 82):
My cell phone is only a few months old and I'm already looking forward to the not only the A50 series, but what AMD, Qualcomm, and Samsung will do with their own 64 bit designs.

Stay tuned.... very interesting announcements to come out of Las Vegas next month.  
Quoting lightsaber (Reply 82):
It is amazing what volume can do for R&D progress.

Driven by commercialization of a hyper competitive, take no prisoners, ballooning market. I can't recall any point in history with a dynamic like this. It is unprecedented.

Quoting lightsaber (Reply 82):
But I feel old!    I see the interns using smartphones for shortcuts that I just do not consider.

The youth certainly keeps one young(er)... though they never let one forget that one is a "geezer"!  
Quoting lightsaber (Reply 82):
If anyone finds a reason to put a cheap IR camera in a smartphone... that would lay the groundwork to change a dozen industries...

Look at 3 D printing... last year there were only 2 exhibitors in Las Vegas... this year 3 D printing will have its own "tech zone"

Quoting peterjohns (Reply 85):
It may very well be technically probable, but would it save money!!??
Quoting peterjohns (Reply 85):
As an ATCO myself I can see many reasons why decision authorized persons (Pilots?) must be on board an a/c.
- medical emergencies, - diversions , resetting the computers in flight (ever flown an Airbus? - circuit breakers)...
weather avoidance, level change due to turbulance, VFR avoidance when flying into according airspace (law)

Huge gobs of money... it has been detailed in prior threads. As for your other points, existing tech (that is not yet applied) renders them moot. As has been detailed, current airliners are flying with "prehistoric"   gear.

Quoting tim73 (Reply 84):
starting from around 1992-1993.

In general I would concur. However, I think that 2007 was the Sputnik moment. 
Quoting rwessel (Reply 83):
Still, a decade ago who'd have thought we'd have (near) inertial nav systems on our cell phones.

Answer: Very, very few! And we are indeed very near. It is giddy to think that one can get a 9-axis INS module the size of a 25-cent coin for ~$30 retail.

Last year, luddites   where not amused when in reply to one of their disaster scenarios - avionics failure, I said in a jocular answer that all the pax' cell phones were backups
Quoting suseJ772 (Reply 86):
If the ATC person (or someone else responsible in the future OR computers just automatically know to change it), can just change it in the computer directly, then the pilot is still not necessary - which is what we are discussing here.

You intuitively understand what is coming... NextGen ATC has the controller simply entering any ATC instructions directly into the aircraft's FMS.
Nationalism is an infantile disease. It is the measles of mankind. - A. Einstein
 
rcair1
Posts: 1147
Joined: Wed Oct 28, 2009 8:39 pm

RE: Flight Global: When Will We Drop The Pilot?

Sat Jan 04, 2014 9:23 pm

Quoting lightsaber (Reply 82):
I'm floored at the progress. Some of the cell phone technology blew past aerospace standards so quickly that established R&D labs we left explaining how 'cheap' commercial technology passed them by so quickly. It is amazing what volume can do for R&D progress.

Having been in R&D developing cell phones (I worked on the first Google phone - tho I did not work for Google or the brand of the phone" - I can tell you I would run away screaming from any aircraft dependent on the type of R&D hacking that goes on in these phones. There is absolutely no discipline in the OS development processes - days before release we were getting daily releases that broke much of the peripherals in the phone (you must remember, the phone is basically a computer system).

Anyway - the advances in cell phones comes from leverage in price/volume and lack of discipline in RnD.

Quoting rwessel (Reply 83):
A certain level of IR sensitivity is actually natural for many of the sensors used in digital cameras.

CCD's and CMOS are pretty sensitive to IR and the on chip filters pass IR. You can build a passible IR camera if you can diddle the filters.

Quoting rwessel (Reply 83):
That's actively filtered out because more than a bit of clothing is semi-transparent in near-IR, which leads to something of a PR disaster. So there may be some resistance to actively integrating good IR in general devices.


   We do not put IR filters in the cameras because of some problem with transparency of clothes. We do it because it screws up the photography.
rcair1
 
Pihero
Posts: 4318
Joined: Mon Jan 31, 2005 5:11 am

RE: Flight Global: When Will We Drop The Pilot?

Sat Jan 04, 2014 9:55 pm

Quoting suseJ772 (Reply 86):
... in this thread it feels like you are grasping at keeping your job and sounds desperate.

The only desperation, I feel is seeing people so miserably missing the *Wonder of being human* and joyously relinquishing all our human-ness to a *perceived* travesty of so called intelligence.

Quoting suseJ772 (Reply 86):
there is no autopilot than can cope with a simple bounced landing.

It doesn't mean it won't be able to cope with it in the future.

It could have been programmed before, although it's an interesting manœuvre...
But It hasn't.
Another overlook, mayhaps ?

I may be just another pilot in love with flying, but I'm also someone who has been working with computers for the past 30 years ( You may say that my experience covers all aviation progress since the DC-4 °.
I am also quite proficient with computer aided decison programs (with three very Advanced Airlines in this respect ).
I could again quote Ms Cummings, but I know for a fact that an *operator + the computer* is a lot more efficient than a computer on its own ( She has a meter at 20 % increased efficiency )... and computing power has nothing to do with it ... the algorythms have and they haven't progressed a lot for the past fifteen years in spite of continuing research by an elite team of ten or fifteen engineers.
One of the problems can be very simply stated : " After an event - be it weather, ATC strike or industrial action related.. - has halted the airline operations at the main hub, how to restart a normal schedule as quickly, efficiently, economically as possible while insuring that the impact on affected passengers is minimal ?"

... and no, I'm not a programmer . The last program I designed was five years ago :a NAM spreadsheet to help me validate fuel planning and another one to give me the headings between two oceanic waypoints, again in order to monitor my navigation. That on a Palm handheld PDA.
... and I'm no gamer as I play with simulators and that's enough for me.

[Edited 2014-01-04 14:01:10]
Contrail designer
 
ncelhr
Posts: 258
Joined: Tue Jul 04, 2006 10:53 pm

RE: Flight Global: When Will We Drop The Pilot?

Sat Jan 04, 2014 11:29 pm

8 years ago I worked on the pilotless aircraft scenario for one of the major aircraft manufacturers.
We concluded that a pilotless aircraft was absolutely possible but that it would take time for pax to adapt to the idea of being in a pilotless aircraft.

There was an old joke going around: rather than having 2 pilots, you could save costs by having a pilot and his dog.
What does the dog do? He wakes up the pilot when he needs feeding - but he also stops the pilot from touching the controls.
What does the pilot do? He feeds the dog.
What about the passengers? They're happy because they know there's a pilot up front.
 

Apologies to all the pilots here. Sick aircraft industry joke.
 
planemaker
Topic Author
Posts: 5411
Joined: Mon Aug 25, 2003 12:53 pm

RE: Flight Global: When Will We Drop The Pilot?

Sat Jan 04, 2014 11:39 pm

Quoting ncelhr (Reply 91):
8 years ago I worked on the pilotless aircraft scenario for one of the major aircraft manufacturers.
We concluded that a pilotless aircraft was absolutely possible but that it would take time for pax to adapt to the idea of being in a pilotless aircraft.

Old joke... but sick???? Naw!  

Acceptance will be a generational think. As pointed earlier, in the prof's class she says that all students put up their hands when asked if they would fly in a UAV airliner. Just like some businesses kept elevator operators on long after they were needed, the same will happen with airlines.
Nationalism is an infantile disease. It is the measles of mankind. - A. Einstein
 
User avatar
SuseJ772
Posts: 935
Joined: Wed Aug 10, 2005 11:13 am

RE: Flight Global: When Will We Drop The Pilot?

Sat Jan 04, 2014 11:51 pm

Quoting Pihero (Reply 90):
I may be just another pilot in love with flying,

Well out of all the people on this board, I would never say you are just another pilot. Your insight into AF447, and other aviation related topics have been greatly appreciated. You are one of the most intelligent pilots I have ever "met." I too love flying (admittedly, on a 172 or an SR-22), but you and I disagree on the business and scientific realities that will determine what happens in this mode of transportation.

Quoting Pihero (Reply 90):
The only desperation, I feel is seeing people so miserably missing the *Wonder of being human* and joyously relinquishing all our human-ness to a *perceived* travesty of so called intelligence.

There were probably a lot of Travel Agents in 1997 that would have said the same thing   

Quoting Pihero (Reply 90):
... and I'm no gamer as I play with simulators and that's enough for me.

Nor am I. Never got the point of it.
Currently at PIE, requesting FWA >> >>
 
Pihero
Posts: 4318
Joined: Mon Jan 31, 2005 5:11 am

RE: Flight Global: When Will We Drop The Pilot?

Sun Jan 05, 2014 12:37 am

Quoting suseJ772 (Reply 93):
you and I disagree on the business and scientific realities that will determine what happens in this mode of transportation.

First, thank you for your kind words   

It's really very simple :

I am quite open. (It's true !)

Where is the *scientific* basis for the proponents of all automation, or even the *remotely monitored unmanned airliner* ?

So far, except for déclarations on their confidence that "in the future...blah blah ....blah " and the super exponential growth of computational power, NONE of my arguments have been addressed.
...and when I cite Ms Cummings - who certainly cannot betaken for a fan girl of the piloted vehicle - and her concerns, surprise ! curtains ! let's talk about gadgets !

May I remind you ofthe subject of this thread ?
And may I remind you ofa little real life situation I ask them to solve, therefore easily code... on which, for the past two years theres been NO answer.

Which as a matter of fact makes me laugh because with such a response, thank you very much, my job is very secure...

One sentence on my above post, quoted from that M.I.T team deserves to be reiterated :
..."even highly automated systems can be presented with a set of dynamic and unexpected variables states never envisioned in advance by designers, which can ultimately lead to catastrophe.

I'd add *unknown* to *unexpected* and IMO *unexpected* means *unforeseen*.

Until that day when wez see human-like intelligence in automatons, this subject is just a day dream.

...and I'm also very polite.
Contrail designer
 
Pihero
Posts: 4318
Joined: Mon Jan 31, 2005 5:11 am

RE: Flight Global: When Will We Drop The Pilot?

Sun Jan 05, 2014 12:45 am

Quoting planemaker (Reply 88):
Stay tuned.... very interesting announcements to come out of Las Vegas next month.

Do you read others'posts or are you just having a monologue ?

If I see well, my post precedes your elegant one-liner by some 14 minutes... and I was just back from Venise and make a kisser on arrival followuing an all-manual approach... the first officer was very pretty and the senior purser didn't intend to steal my job as a commander of that flight...

Pure bliss !

Puts things into perspective, doesn't it ?
Contrail designer
 
lastrow
Posts: 70
Joined: Thu Dec 30, 2010 10:33 pm

RE: Flight Global: When Will We Drop The Pilot?

Sun Jan 05, 2014 12:47 am

Interesting thread, thank you all!

Quoting planemaker (Thread starter):
What he is describing is doable now. By 2030 information technology will be 1,000's of times more powerful.
Quoting planemaker (Reply 7):
The technology is already basically "developed". If we don't have UAV airliners by 2030 (be they RJs or NBs) it certainly won't be because the tech isn't developed.
Quoting Rara (Reply 18):
The basic technology is already available, and doubtlessly will only improve in the years to come.

While nobody can foresee the actual required technological challenges to be solved, I think most are on the wrong path when it comes to "available" technology: Some argue with intelligence and CPU capability, but to my understanding there is no robust communication channel so far to communicate with the aircraft in case of problems. You could assume that the plane acts autonomously in all situations.

If that is not the assumption, a robust communications link must be developed. There is technology today for providing real-time (=reliable) indoor communications. However, atmosphere is a different medium. Someone has mentioned communication lag, which comes also in place. My bet is that computational intelligence won't be a problem, but reliable communications to the plane will be.

Quoting FlyPNS1 (Reply 25):
The technology for automated rail lines has been around since the 1960's. Yet even today, almost every major subway and rail line still has engineers or drivers.

Because investments for subway lines are long term. It does not make much sense to update existing lines. Where new subway lines are built, operator-less cars are viable options - and installed, as seen in many cities now.

Quoting lightsaber (Reply 82):
Some of the cell phone technology blew past aerospace standards so quickly that established R&D labs we left explaining how 'cheap' commercial technology passed them by so quickly. It is amazing what volume can do for R&D progress.

such technology is by far not suitable for aircraft operations. You should know that even for microprocessors you cannot use consumer grade chips but need hardened / certified versions of designs. Software in such devices offers maybe "five-nines", if not three. If your smart phone does not work properly you are advised to restart it. One solutions could be to get as many different redundant components that you achieve the reliability by statistics - which is the approach of real time wireless communications which is limiting the bandwidth, however.

Quoting planemaker (Reply 27):
Robot cars are ALREADY approved for use in some states under certain conditions. Google's fleet of robot cars have driven close to 1,000,000 miles all over California.

OK, try to crash drive such car into some building and see what happens. Then, please explain how you would like to compare this to flying planes. To give another example: there are vacuum cleaner robots that autonomously navigate through the apartment and reliably clean the floor. So the argumentation is that because we have this available now, autonomous planes are in sight?

Quoting rcair1 (Reply 89):
I can tell you I would run away screaming from any aircraft dependent on the type of R&D hacking that goes on in these phones.

  
However, maybe are Linux-based devices certified for some higher SIL (I do not know a source, but that is my guess).

Quoting suseJ772 (Reply 86):
Which of course is now owned by Google.

Google bought the mobile phone business of Moto. The chips that were mentioned here went into Freescale today.

Quoting lightsaber (Reply 82):
The neat thing about automation is once it is installed, software upgrades will allow improvement in efficiency, safety, timeliness, information transfer, durability, and diagnosis.

And the nasty thing is that it will require re-certification. Safety critical systems cannot be upgraded as with your auto-update on your smart phone.

Quoting tim73 (Reply 84):
Maybe those processes are a bit outdated?

The point with certification is that there are huge efforts in order to build systems eligible for certification. Improvements there (automated formal analyses, automated toolchains, automated traceability, etc) will open new doors there.

Quoting planemaker (Reply 79):
I've seen it quoted a few times that a $2 musical greeting card has more computing power than the Apollo 11 Lunar Lander... and we toss the card in the garbage.

The ICs on the musical greeting card will likely not work on the moon (temperature range, radiation, vibrations...). Such comparison with available consumer technology is not useful.

Quoting lightsaber (Reply 61):
But that mac I used only had a 16Mhz processor, but it was a 68030 baby!    And it had a million bits of RAM! (128k) that was upgraded to a whole 4 million bits of ram (512K) by 1992. Anyone remember the first Sim City?

offtopic, but the IIcx would not have booted with 128kb of RAM. My guess is that your machine had one Megabytes and you upgraded this to four MB.

Quoting lightsaber (Reply 65):
The 68030 has a pair of 256 bite cashes. Yes, not kilo or mega bite size, but bites..

offtopic, but apart from that it is "cache" and megabit and megabytes, there was also a reason for that: the IO was clocked at the same speed as the proc clock and the wafer yield was lot lower and thus die size was expensive (think of the pentium pro specs) so huge on-die caches did not pay off compared to external (off-die) caches. The successor of your cited IIcx had some 32kb cache SIMM installed optional, it was the IIci.

Quoting wowpeter (Reply 57):
So what happen if all those fails you and something like this did indeed happen? It is a possibility, something that we need to do to think outside the box. We can't just say not possible and be done with it

The uncontained engine failure that you have cited is regarded possible. However, weight restrictions on planes are the trade-off considered here. it is just not economic to build a shielding that prevents that. Power turbines of the same output installed in power plants weight ten times as much.

Quoting tim73 (Reply 55):
Space shuttle codebase: 40 000 lines of code.
US military drone, control: 3.5 million
Firefox browser: 9 million
Windows NT 1996: 12 million
Boeing 787, avionics only: 6.5 million
Boeing 787, total: 14 million lines
Linux 3.1: 15 million
F-35 Fighter: 24 million
Windows 7: 39 million
Large Hadron Collider: 50 million
and then....average modern high end car: 100 million!

this is truly an apples with oranges comparison. OSes are general purpose software aggregates that carry a lot of functionality, if you would tailor OSes to the actual code base that is needed by a particular user it would be much less. Moreover, do you know at what reliability PC software is delivered today? Today desktop software is shipped with about 1% of the tests failing (if the company has some decent test infrastructure), the 1% will be corrected in the next release.

To get back to the topic: I agree with those who have projected a dedicated short hop route that is freed for drone-like traffic for cargo as the first application.

A thing that has not been mentioned here, is that space missions are in most cases "pilot-less" and it was also argued by some that the manned Rockwell space shuttle was a design "disadvantage" that has risen the cost of the entire programme unnecessarily (as opposed to the unmanned Buran design). But as with military application we have lot lower reliability here. I think space shuttle missions have a 1:50 failure rate ...

And for those pilots who fear their jobs wont be safe: there will be still a high number of conventional planes that are paid, build and require pilots.

looking forward to your posts  , lr.
 
tim73
Posts: 93
Joined: Sun Nov 17, 2013 9:03 pm

RE: Flight Global: When Will We Drop The Pilot?

Sun Jan 05, 2014 2:14 am

They say about France:
You must act different but you must think the same!
And Germany:
You must act the same but must think the different!
 
User avatar
lightsaber
Moderator
Posts: 18672
Joined: Wed Jan 19, 2005 10:55 pm

RE: Flight Global: When Will We Drop The Pilot?

Sun Jan 05, 2014 4:20 am

Quoting rwessel (Reply 83):
That's actively filtered out because more than a bit of clothing is semi-transparent in near-IR, which leads to something of a PR disaster. So there may be some resistance to actively integrating good IR in general devices.

Ok, you just sold that phone to the male half of the population.   Forget bad PR, I'm seeing this is a strategy for BadAss Vodka or another edgy brand... We just found the market niche to get really cheap IR cameras quickly.   

Quoting rwessel (Reply 83):
The further from the visible spectrum you get the more expensive things become, although again, a fair bit of that is volume related.

Volume is where it is at. Aerospace will never have volume and thus must ride on other technologies. But I've certified commercial hardware for aircraft use to take advantages of the economies of scale. The issue is keeping vendors from changing configuration or manufacturing process. One option is UTC's approach for processors... buy in batches so large its justified to certify each batch *and* demand rights to buy the last stepping of the chips at the last production process before product switch. Yes, they have to pay for those contract provisions... And yes, the foundries have to be part of the contract too...

Quoting rwessel (Reply 83):
Still, a decade ago who'd have thought we'd have (near) inertial nav systems on our cell phones.

The inertial systems in some game controllers are better than many flight controllers of a decade ago, even better than what is in the cell phones (at least this month...). The vendors have been approached by aerospace companies and *paid* to certify parts, in particular for some of the smaller drones.

Quoting tim73 (Reply 84):
I am pretty sure, the cell network&phone R&D has been the number one technology when it comes to money spending, starting from around 1992-1993. It required so many breakthrough innovations via expensive research projects. No wonder the fruits of those projects are now spreading to many other areas too.

   Just amazing. While some of the work isn't pretty... Enough is... and the code bases are getting cleaned up quickly thanks to the need for rapid software patches that work the first time.

Quoting tim73 (Reply 84):
It is kind of funny, any DYI amateur could build a better small surveillance drone with things like Nokia 41 megapixel monster Pureview 808 optics

What amazes me more is I bought a $19 toy helicopter with a flight stabilization system so advanced a 5 year old can fly it... A decade ago the first gyros were made for RC aircraft and those are nothing compared to what can be done now. Seriously, a $19 (ok, its $20.50 today... I guess I bought it on a cheap day...)

http://www.amazon.com/gp/product/B00..._details_o00_s00_i00?ie=UTF8&psc=1

Quoting suseJ772 (Reply 86):
I'd like to thank Lightsaber and Planemaker for all the information in this. As a CTO and a private pilot, I found your insights very interesting.

Thank you. But do realize I have learned a tremendous amount on this thread. Some from posts here, some to research to answer questions or challenges. This has been one of the best threads in a while.

I thank Planemaker. Personally, I think he deserves a hattip for helping direct this thread on such an informative course.

Quoting planemaker (Reply 88):
Stay tuned.... very interesting announcements to come out of Las Vegas next month.

   Yes, I'm following the news... I find it interesting how a bunch of auto makers are making announcements to include cell phone technology into their cars... See below on my comments on their certification processes.  
Quoting planemaker (Reply 88):
Driven by commercialization of a hyper competitive, take no prisoners, ballooning market. I can't recall any point in history with a dynamic like this. It is unprecedented.

It is. And it brought 'pervasive computing' to the world several decades ahead of schedule.

Quoting planemaker (Reply 88):
Look at 3 D printing... last year there were only 2 exhibitors in Las Vegas... this year 3 D printing will have its own "tech zone"

3D printing will change economy of scale forever. Now old products don't have to go out of production, just raise the price to what 3-D printing can make the part at on a one off custom ordering basis.

Quoting planemaker (Reply 88):
It is giddy to think that one can get a 9-axis INS module the size of a 25-cent coin for ~$30 retail.

   Who saw that coming? I didn't and I was telling everyone how certain technologies would filter down... Yet the pace was *far* faster than I imagined. I watched stuff happen in 2012 and 2013 that I was predicting in 2009 would happen by 2020.   

Quoting lastrow (Reply 96):
such technology is by far not suitable for aircraft operations. You should know that even for microprocessors you cannot use consumer grade chips but need hardened / certified versions of designs.

Yes. But hardening is now often in the computer rack, not the chip. Automotive chip certification today is often more stringent than aerospace (due to the volumes required and need to eliminate warrantee expenses). Since automobiles are adopting cell phone technology at a rapid rate (often only a year behind the curve), the delta qualification is usually cheap to not-required! (The usual culprit is the thermal shock test and possibly EMI testing.) There is a reason PowerPC is the current darling of aerospace processors, FORD did such a good certification its like Christmas...    It really is a different world when one is worried about warranted cost on tens of millions versus hundreds of parts...

Now if the automakers would adopt aerospace thermal shock and pressure transients.   (Hint, they won't as they don't need to pay for those tests.)

Quoting lastrow (Reply 96):
And the nasty thing is that it will require re-certification. Safety critical systems cannot be upgraded as with your auto-update on your smart phone.

Agreed and since I posted about my time in a system center lab, you would know I have seen aircraft software certification and the process. Recall in aerospace delta qualification of exiting parts is more often the norm rather than a full re-qualification. We're teaching other businesses who are getting into the aerospace market how to do it. I've seen aircraft software certification processes timelines cut by an order of magnitude since 2007 on two different programs. Let's just say the cost to build and maintain a full ironbird/system center lab is worth it. Then again, I'd like to get back into such a lab (but for new, not established aircraft... I'd get bored doing maintenance builds after EIS...).

Naturally, some vendors have such poor qualification processes that they cannot be used for aerospace... But perhaps laid off aerospace engineers moved down from LA county to the Irvine chip development companies that some are good enough and happy to sell, at quite a mark up, to their old employers.   Not that I could prove any examples...   

I've taken $30 commercial parts, applied a $370 X-ray and acceptance test procedure program to them and replaced expensive aerospace parts after a qualification analysis and test program. So I know it can be done. The parts won't ever be as cheap as commercial parts... but I did this in under six weeks with the help of only three other people creating and executing the delta qualification plan/analysis.

We didn't even tell the vendor anything other than pay a small premium to receive parts for that same batch and pay them to warehouse a bunch of parts from that batch. As long as the vendor beats their ROI targets, they're happy to do business on such a scheme. In this case, we released the parts back to the vendor after buying only a few shipsets, but they were happy as we had to pre-pay their warehousing and interest costs for far longer than it took them to sell the parts into the commercial market. Win-win. Our total costs were 10% of the total costs of an aerospace part and if it had gone into mass production our total costs would have actually grown relatively; to say 15% of the cost of the aerospace parts) due to economies of scale would have brought down the aerospace costs per part but not the commercial parts due to the need to fully repeat the delta-qual testing on future batches and some costs lining up such as X-ray and material verification.

Quoting suseJ772 (Reply 93):
There were probably a lot of Travel Agents in 1997 that would have said the same thing

I felt guilty defecting to online travel agencies (I really liked my travel agent), but it was so much easier to find a cheap *convenient* flight where I spent an hour going through the options. I couldn't go back.

Quoting lastrow (Reply 96):
To give another example: there are vacuum cleaner robots that autonomously navigate through the apartment and reliably clean the floor. So the argumentation is that because we have this available now, autonomous planes are in sight?

The differential equations programmed into Kiva factory robots are far more complex than those required to fly an aircraft safely. Those techniques and the engineers who work them will filter back into aerospace.

Quoting Pihero (Reply 36):
And I have news for you: there is no autopilot than can cope with a simple bounced landing.

That is false. I worked testing software than was testing landings on windshears that no human pilot could react to and the software passed first time through. A bounce isn't a complicated differential equation to solve. A carrier landing is inherently a multi-bounce landing (not ideally, but potentially)! On a bounce, a human pilot goes around while the UAV will stick to the deck. Heck, system software on UAVs is tested under simulated windshear conditions that have crashed piloted planes. The Lion air 738 crash would have been avoided if a Grade 2 autonomous system was installed (but it would take an aircraft more advanced than the 737 or A320 to have the control surfaces react as UAV control surfaces react).

Computers are also better at holding in tight stall tolerance, in particular during turbulence. Now this won't be as useful in commercial aircraft due to the very conservative mandated minimum climb rates, but it does allow much more advanced laminar flow wings to be implemented on commercial aircraft. I'm actually not certain if some of the best wing aerodynamics could be implemented without some autonomous backup... I'm curious. I'll have to do some research.

I do not argue we're ready for Grade 4 automation today nor socially ready even in two decades. But it will be a nice safety move to offer Grade 2 or 3 automation (it would cut insurance rates) and eventually go to solo pilot operations with the backup pilot in a control center. e.g., the Helios 522 passengers who donned their oxygen masks could have been saved if there had been a remote way to land or autonomous way to land.

We're at the point where autonomous aircraft flight test programs now go much quicker than piloted programs as the computer will do as you command it instead of 'this is how its always done.'

Quoting lastrow (Reply 96):
but the IIcx would not have booted with 128kb of RAM. My guess is that your machine had one Megabytes and you upgraded this to four MB.

Fair enough. It was the roommates computer.   I just remember he went from the minimum ram of '1 unit' to 4units. I couldn't afford a Mac IIcx back then... It wasn't until 3 years later I finally broke down and bought my first computer instead of using borrowed, school, or parental 'hand me downs.' A computer that needed a better video card within a year (back when every six months the speed doubled), and a new motherboard & CPU, and video card within two years... To play a flight simulator called realflight.    How I date myself...   

Lightsaber
IM messages to mods on warnings and bans will be ignored and nasty ones will result in a ban.
 
lastrow
Posts: 70
Joined: Thu Dec 30, 2010 10:33 pm

RE: Flight Global: When Will We Drop The Pilot?

Sun Jan 05, 2014 9:47 am

Quoting lightsaber (Reply 98):
The differential equations programmed into Kiva factory robots are far more complex than those required to fly an aircraft safely. Those techniques and the engineers who work them will filter back into aerospace.

I see, logistics for factories is a more advanced example than vacuum cleaning. ... but, if you have heavier robots in factories you have this a no-go area for humans, AFAIK for safety. and still workers are heavily injured from time to time because of accidents with robots.

Quoting lightsaber (Reply 98):
That is false. I worked testing software than was testing landings on windshears that no human pilot could react to and the software passed first time through. A bounce isn't a complicated differential equation to solve.

as said, I agree fully with you that computational complexity for decision making when flying a plane appears possible. I am just not agreeing with the optimism to make it actually happen, on planes because the challenge is actually to bring it to the domain / safety level. (and in addition I do not see the communications link solved by now)

Also control theory brings me to exactly this thought: There will be many inputs at the same time that a control needs to output for apart from reacting to wind shear: (faulty) engine behavior, obstacles on the runway, etc.

Currently, individual tasks can be easily solved with a magnitude of higher precision by autonomous systems. We agree on this, even bounced landings.   The fun lies in putting different parts together and model their interaction or let's say here: interference. And then get SIL 3 (or 4?) certification with proving no interference will result in harmful behavior. Please, do not get me wrong. Surely flying will be autonomous at some time. A am just not optimistic it will happen so easily because of seeing what we have now.

To conclude: A recent technology step appeared difficult to get into regular operations: the bleedless engine resulting in a different electrical architecture on the plane - not new functionality provided by machines, just making it better by a different way (what ever better means=more efficient, less costly, ...). Compared to autonomous flying the bleedless engine appears as small change ...

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