Which would be more dangerous?

[complexintentions]

[

Point being, the automatics are only as good as the programming. Until that changes - no flight without some sort of human intervention - no matter how many times removed.

I agree but the issue is "does the human have to be in the airplane" . Autonomous aircraft are a long way off, but I think airliners that have that human intervention delivered, when required, by a data link from a person in a room full of computer screens in some industrial park is a lot closer than you think...

I have never disagreed with that per se. That's why I broke the argument in two parts.

However, completely unmanned flight, even with remote intervention, doesn't seem like it's too close either. When I think of the polar ops we do where there is no comms whatsoever, the ops in monsoon weather and so on, where telemetry could so easily be disrupted - I just don't see it working. I mean, the sat phone and onboard wi-fi doesn't work - but the plane will be flown by someone back in the US or whatever? Doubtful, at this point in time and tech. And to build it to the spec that might meet the required level of reliability - something like might be used in military applications - would seem to be prohibitively expensive.

I can't imagine how many redundant systems there are to guide a Predator to its target. But the US Navy doesn't have to turn a profit.

[Big Pistons Forever]

When the first 777 freighter takes off with no pilots on board the airline pilot profession as we currently know it, is over. I think that event will happen in less than 20 years from today, possibly in as little as 10 years, although I admit that is probably a bit ambitious.

[DonutHole]

Except we still can't make a computer that will Never crash, or sensors that won't fail

That's true. That's why you mitigate risk by using monitoring systems and system redundancy.

You're in the vacuum. What happens when the cell phone drops the call. A computer only can do what it is told, it can't make decisions... When shit hits the fan it beeps and asks for a human to fix it. We are the redundancy. He's sitting at anothe machine, plugged into a grid, with a redundant generator. Everybody working off navaids which go down too.

You're not talking about an aircraft with a computer and some servos. You're talking about relying on levels upon levels of machines and computers and redundancy. It's just too expensive... we will be constantly upgrading an rethinking this thing as it evolves.

Having humans at the controls will always be cheaper, and safer when the it goes down or the data link fails. We can't even make bulletproof Internet.

Please continue to show me how imperfect beings can create a perfect machine.

I don't think I ever used the word "perfect" in describing automation in aircraft. It's all about risk management and humans have proven to be much more variable when it comes to completing a task. So, can you show me a "perfect" human?

http://avherald.com/h?article=483602c2&opt=0[/quote]


Good thing there was a "spare" on board.

and the shuttle doesn't land itself... for the record.

That's true. The easiest procedure in the entire mission is left to the human pilot 25 miles from the landing strip. I don't want to minimize the activity but it was always done in good weather conditions albeit "dead-stick" at with very high approach and landing speeds. Generally, auto-land is reserved for low visibility.[/quote]

it will take one big accident and this pipe dream will be over.

[Chuck Ellsworth]

Having humans at the controls will always be cheaper,

Watching the gradual downward spiral of pilot wages with no bottom in sight is proof positive of the above.

[lownslow]

So I wasn't there but I wonder if the same conversation was happening when it was thought that navigators would be redundant, or when fly by wire was being introduced?

Maybe I should clarify the original question: assuming pilotless aircraft will happen some day, will (or should) the crew number go from two to one to zero, or from two straight to zero?

[Big Pistons Forever]

Maybe I should clarify the original question: assuming pilotless aircraft will happen some day, will (or should) the crew number go from two to one to zero, or from two straight to zero?

I think you will see the following progression

-Freighter: One crew
-Pax: One crew
-Freighter: no pilot selected trans oceanic routes
-Pax: no pilot selected trans oceanic routes
-Freighter: gradual route expansion including some domestic routes
-Pax: gradual route expansion including some domestic routes
- Freighter: Most routes
- Pax : Most routes

I can see some routes never going crewless due to lack of landing aids, special procedures required, remote location etc etc

I think the the concept will be proven on freighters and then migrate to pax ops and will move from the least complex operationally to the more complex. I predict that SFO & LAX to China will be the first. Departure and arrival routings in the US airports can be kept almost entirely over water and the Chinese population will be told what the government wants with no choice in the matter.

[Rockie]

I alluded to this earlier, but the Future of Life Institute recently published an open letter on the risks and benefits of developing artificial intelligence.

http://futureoflife.org/static/data/doc ... rities.pdf

It is safe to say that the signatories of this letter have devoted more thought to this topic than we have here at AvCanada and are infinitely more qualified than we are. They include - but are certainly not limited to - Elon Musk (who just announced an autopilot system for the Tesla) Bill Gates and Stephen Hawking. In discussing automated vehicles controlled by AI they go no further than automobiles which can at the very least be equipped with a big red button on the dash that says "STOP" so that if something goes wrong it will, well...stop. Can anyone here imagine a scenario where an airplane in flight can just "STOP"? Has anyone noticed that the red toggled emergency stop switches found in the simulator don't exist in the real airplanes?

The letter is a far ranging document that delves into many areas of concern that you can read for yourself, however the ones that caught my eye were security and control. Here is what they say about control:

2.3.4 Control

For certain types of safety-critical AI systems - especially vehicles and weapons platforms - it may be desirable to retain some form of meaningful human control, whether this means a human in the loop, on the loop, or some other protocol. In any of these cases, there will be technical work needed in order to ensure that meaningful human control is maintained. Automated vehicles are a test-bed for effective control-granting techniques. The design of systems and protocols for transition between automated navigation and human control is a promising area for further research. Such issues also motivate broader research on how to optimally allocate tasks within human/computer teams, both for identifying situations where control should be transferred, and for applying human judgment efficiently to the highest-value decisions.

I was going to underline the key words to highlight them but the whole paragraph is key and bears very careful reading.

[Big Pistons Forever]

It seems to me that the discussion has drifted significantly.

It is a long stretch from no pilots in the aircraft to no pilots anywhere in the loop, or a fire and forget airplane.

A datalink remote pilot like the predator is todays technology and enables remote operation over about 90 % of the worlds airspace. The missing bits of a gate to gate pilotless flight today, is the taxi and takeoff both of which there are existing technological solutions.

The only constraints are passenger acceptance, regulatory approval, and most importantly whether or not the business case is compelling enough.

Time will tell how the pilotless airliner model will play out but to maintain the status quo means that airline flying is exempt from all the technological and commercial pressures that are driving change in every other part of our life.........

[Rockie]

It is a long stretch from no pilots in the aircraft to no pilots anywhere in the loop, or a fire and forget airplane.A datalink remote pilot like the predator is todays technology and enables remote operation over about 90 % of the worlds airspace. The missing bits of a gate to gate pilotless flight today, is the taxi and takeoff both of which there are existing technological solutions.

The original premise of this discussion was removing the pilot from the loop since they are the ones making the mistakes and causing crashes. From that aspect it make no sense to have a pilot still involved but remote, which dramatically decreases his level of integration with the flight and his natural caution. Predators are the worst possible example in the link from now to pilotless flight. They exist precisely because they do not risk a human being (friendly one anyway) and the only loss if one goes down is money.

To eliminate human error you have to eliminate humans from the chain which you cannot really do anyway since humans will build the chain to begin with. Removing their ability to respond and correct errors is a recipe for avoidable disaster. So if humans are going to be involved anyway the most efficient place for them to be is onboard.

[Rockie]

Time will tell how the pilotless airliner model will play out but to maintain the status quo means that airline flying is exempt from all the technological and commercial pressures that are driving change in every other part of our life.........

Since when has any aspect of aviation stuck on the status quo? Advances present themselves continually since flying was invented and the struggle for us as pilots and operators is to keep up. Assuming human removal from the cockpit as the inevitable direction we're moving toward is a mistake though. There is no perceivable upside and far too many downsides.

[CID]

The fatal flaw in your argument is that machines don't even have the "choice" of a decision.

That's a little over-dramatic but I see where it's coming from. I have used the term "risk management" in this discussion several times and it's significant. To suggest that any mode of flight control (human or computer) is going to be flawless is absurd. Having said that, there are several "fatal" flaws in each mode.

Humans, can literally have "fatal" faults. They can die or become incapacitated due to sickness or lack of oxygen. As far as "fuzzy logic" goes, I don't think you're using the term correctly but I understand what you meant. And...if you are suggesting that humans are the masters of such logic and decision making one only needs to look at airline accident causes to see that humans often make the wrong conscious decisions and often those decisions prove to be fatal.

What a computer will generally NOT do is deviate from the "program". The program or "algorithm" can be very complex to take several variables into consideration. Think of the warning algorithms in TAWS computers. They essentially take many parameters into account to determine if the pilot needs to take evasive action to avoid hitting the ground. No AI required.

That's right Rockie.....no AI, no conscious computers. This is all simple computing that is well within current technological capabilities. In my opinion (and several others in the field) automation is not meant to mimic human behaviour and human nature. It's meant to take the variability out and reduce workload. The term "reduce workload" is a polite way of saying, this process is too complex for a human to consistently reproduce positive results or we can take another human out of the cockpit. Think of the 747-400. Automation basically took the flight engineer position away.

To eliminate human error you have to eliminate humans from the chain which you cannot really do anyway since humans will build the chain to begin with. Removing their ability to respond and correct errors is a recipe for avoidable disaster. So if humans are going to be involved anyway the most efficient place for them to be is onboard.

That depends on the process you are talking about. The beauty of computers, algorithms and databases is that they can be analyzed and corrected before putting them into service where they perform the exact functions repeatedly with no deviations. You can't tell me that every pilot that comes out of training will fly the airplane exactly the same as every other pilot. Some pilots are good, some are bad, some are a little too "brave" and some are a little too "chicken". Some are masters and running the FMC/FMS and some let the kid in the right seat do it. Computers that are certified to do an automated task have to meet specific design and certification standards and the first one will be the same as the next one and the next one. And continued updates will only make it better.

Like I said before though, this is one of those topics where there are strong divisions. The romantic pilot will never accept the level of automation that will take him out of the cockpit and the public is largely on your side. Your arguments however (Rockie) are pretty weak.

[Rockie]

What a computer will generally NOT do is deviate from the "program".

Precisely their fatal flaw. If any given very specific scenario isn't part of their program they stop and can go no further. You think we can program every scenario in when that is simply impossible - new ones are being encountered everyday. And the strength of humans is they can apply knowledge learned from other events to the new ones, recognizing an evolving scenario early and modifying their actions to prevent potential problems. Non-AI programming will never be able to adapt to a new situation and airplanes would be crashing everywhere.

[Driving Rain]

http://www.faa.gov/news/updates/?newsId=82225

It's starting, although I suspect if successful it would reduce the amount of delivery vans on our roads.

[Big Pistons Forever]

. There is no perceivable upside and far too many downsides.

The problem I find with most of your arguments remains that you are looking at the problem as a pilot. You are absolutely correct there is no upside to you and plenty of downsides to you, but at the end of the day what you want is irrelevant.

From the airline managers view the upside is obvious. Pilots are on the cost side of the ledger. As soon as a pilotless solution can demonstrate less cost than the existing crewing model than there is going to be tremendous pressure on regulators to approve it. As for downsides well all industries have shown an appetite to take risks, often very significant risks in the pursuit of short term gain.

Even when those risks when horribly wrong they generally managed to duck and cover. I made a tidy profit from BP shares I bought 2 weeks after the catastrophic spill in the Gulf of Mexico. BP is still alive and kicking and they were front and center for one of probably the top five business f*uckups in history.

In any case as pilots the time to get worried is when somebody gets approval to fly with only one pilot. The "one pilot" model is the thin edge of the wedge as it is clearly going to be the stepping stone to no pilots.

[Rockie]

You are absolutely correct there is no upside to you and plenty of downsides to you, but at the end of the day what you want is irrelevant.

Actually there are no upsides and plenty of downsides for future passengers - which is the perspective I am seeing this from. None of this will affect me or anybody I currently know.

And only the most short-sighted management would view pilots only from the cost side of the ledger. Every time one of us pulls a rabbit out of the hat - and there are many, many examples - management falls all over themselves bragging about it.

[Shiny Side Up]

Actually there are no upsides and plenty of downsides for future passengers - which is the perspective I am seeing this from. None of this will affect me or anybody I currently know.

There doesn't need to be upsides for the passengers, businesses need only to be able to sell these things as upsides. It could easily be touted as safer and more cost effective and at the end of it if it means a lower ticket price, the average consumer will be sold. Which is the worst part, pilotless airplanes don't have to be "safer" than piloted ones, only "as safe as" to be sold to the ticket buyer.

[Rockie]

It could easily be touted as safer and more cost effective

That won't fool anybody with a home computer...

[Big Pistons Forever]

And only the most short-sighted management would view pilots only from the cost side of the ledger.

Well sorry to burst your bubble but every senior manager at all the airlines view their pilots from the cost side of the ledger. Something that is evident from their actions towards the airlines pilots, which in virtually every instance at every airline has been moves to reduce the cost to the company of employing the pilot group. It is also very evident in the language in their annual prospectus documents and public statements.

Minimizing costs is just good business, which is why as I stated earlier, when the total costs of a airliner flight will be less with no pilots, it is going to happen. The question is when, not if.

My vote is the first pilotless airliner flight with passengers on board will happen in 20 years or less........

[photofly]

You are absolutely correct there is no upside to you and plenty of downsides to you, but at the end of the day what you want is irrelevant.

Actually there are no upsides and plenty of downsides for future passengers - which is the perspective I am seeing this from.

$50 per ticket cheaper, and no flights lost to stroppy pilot industrial action - two huge upsides for passengers.

And only the most short-sighted management would view pilots only from the cost side of the ledger. Every time one of us pulls a rabbit out of the hat - and there are many, many examples - management falls all over themselves bragging about it.

Uh, no. Airline management loathes hero pilots with a passion. Sully never flew again, and the guy that saved the pax in the 777 crash landing at Heathrow was out of work for years. No airline wanted to touch him. Bragging about it falls into making the best of a bad job mostly driven by sensationalist media on the hunt for the next "hero". However every hero pilot makes the rest of the line pilots look weak. Who wants to fly if your fate depends on having a hero up front to save you? Passengers want to - have to - believe that every safe flight is humdrum, trivial, unexceptional. Prime target for automation - and that's what airline management works overtime to feed to them.

[Rockie]

Passengers rightly or wrongly are comforted by a pilot with grey hair over one who looks like their kid - you really expect them to accept no pilot?

$50 per ticket cheaper, and no flights lost to stroppy pilot industrial action - two huge upsides for passengers

$50 you pulled out of your ass and has no basis in reality. Neither does your opinion people will give a single moment's thought getting on a pilotless airplane to the trade off for some future cancelled flight due to pilot action.

My vote is the first pilotless airliner flight with passengers on board will happen in 20 years or less........

I'll take that bet...name the amount.

[Shiny Side Up]

It could easily be touted as safer and more cost effective

That won't fool anybody with a home computer...

Then you have a much higher opinion of the general public than I do. Keep in mind that from same said general public, guys like Creflo Dollar - as mentioned in another thread - can make a fairly luxurious living. These days the general public fall prey to the dubious value of the weekly miracle food or drug, they elect a steady stream of officials who operate not in their best interests. People still smoke, drive while texting, and fly on asian airlines. Ryan air fills seats based on "cheap flights" alone. Given the decreasing quality of pilots, I don't think it will be a hard sell. Just need a few more Wun Wing Lo and Wi Tu Los at the controls.

[Rockie]

It could easily be touted as safer and more cost effective

That won't fool anybody with a home computer...

Then you have a much higher opinion of the general public than I do. Keep in mind that from same said general public, guys like Creflo Dollar - as mentioned in another thread - can make a fairly luxurious living. These days the general public fall prey to the dubious value of the weekly miracle food or drug, they elect a steady stream of officials who operate not in their best interests. People still smoke, drive while texting, and fly on asian airlines. Ryan air fills seats based on "cheap flights" alone. Given the decreasing quality of pilots, I don't think it will be a hard sell. Just need a few more Wun Wing Lo and Wi Tu Los at the controls.

I agree we're complete suckers to clever shysters, but people who don't know much about computers are naturally distrustful and a little intimidated by them. People who know lots about computers know their limits as opposed to a human brain. Neither will knowingly place their lives at the sole digitized mercy of one in an airplane without a pilot. Nor should they. Especially when you have people who really know what they're doing telling them they shouldn't.

[Shiny Side Up]

People who know lots about computers know their limits as opposed to a human brain. Neither will knowingly place their lives at the sole digitized mercy of one

News: they already do. Have you seen car commercials these days? Cars park themselves, and now can automatically take collision avoidance. Given the grotesque amount of risk to the average human being, trusting to such technology seems an awful leap, but the general public seems to have embraced it fully. Now to me as someone who takes a bit of pride in maintaining a skill (driving and flying) the idea of a computer taking over for me is, like you, a repugnant thought, but you must remember that its not to the majority of the population who will gladly surrender the responsibility.

One sees this frequently in the small airplane world as well, I run into lots of guys who would gladly have a private machine that could do everything itself. The sooner they can engage auto, the happier they are. I'd bet there's a market for a private plane that doesn't need a pilot, you can just hop in, say "fly me to Vancouver" and it does the rest. You read the paper and catch up on your emails. All the perks of having a private airplane with none of the pesky maintaining a skill. I think there's the potential that's where we might see this first. Could you imagine if Cirrus started making drones? Only training you need is if something goes wrong you call OnStar and hit the big red button for the chute.

[Rockie]

People who know lots about computers know their limits as opposed to a human brain. Neither will knowingly place their lives at the sole digitized mercy of one

News: they already do. Have you seen car commercials these days? Cars park themselves, and now can automatically take collision avoidance. Given the grotesque amount of risk to the average human being, trusting to such technology seems an awful leap, but the general public seems to have embraced it fully.

What risk are you talking about? Even if people ran the risk of high speed flaming death while a car parks itself or avoids running into another car, the driver is sitting right there to take over at any time. Sorry, but I don't come close to seeing a parallel...

[CID]

People who know lots about computers know their limits as opposed to a human brain.

Maybe. But are you really equating your laptop with computers used for critical functions in airplanes? Day and night. Apples and oranges. Black and white.