Boeing studies pilotless airplane

[Boreas]

Think about it this way, a 172 only requires one engine yet 2 engines is safer. Why? well at a point the risk reward doesn't add up, at the end of the day safety costs money and the line usually gets drawn around the 9 passengers in the back. Then it becomes more essential to have the added layer of safety.

on a side note their is wayyyyy more movements of airliners every day than GA aircraft, yet GA aircraft have a much higher rate of casualty so keep in mind how safe things are right now.

Except that statistically, two engines are not much safer than one. Twin engine piston aircraft have a similar fatal accident rate to single engine piston aircraft of the same class (i.e. faster, traveling singles), and two engine turbo-props have a similar fatal accident rate to single engine turbo-props of the same class. I don't think there's much data for comparing single engine cabin class jets since there's so few around.

Interestingly, this article talks of a study comparing single pilot and two-pilot ops in biz-jets. Same aircraft, only difference being the number of crew members. The sample size is low, but the fatal accident rate is comparable; close enough that they stated it is inconclusive.
https://www.ainonline.com/aviation-news ... -advantage

I don't think that comparing fatal accident rates is a good way of looking at single vs. dual pilot operations.

There are a lot of things that have to go wrong before people die. Issues such as altitude deviations and runway/taxiway incursions are much more common in single pilot operations - especially if operating high performance aircraft in congested areas. These issues, ranging all the way up to non-fatal accidents, would not be accounted for by simply examining the fatal accident rates.

[fish4life]

http://m.scmp.com/news/china/society/ar ... pit-window

FO injured / half sucked out, CPT loses all automation which presumably would have cut the ground to flight connection. This just happened and is another example of how 2 up front increases safety yet doesn’t show up in the statistics.

[StudentPilot]

Think about it this way, a 172 only requires one engine yet 2 engines is safer. Why? well at a point the risk reward doesn't add up, at the end of the day safety costs money and the line usually gets drawn around the 9 passengers in the back. Then it becomes more essential to have the added layer of safety.

What/who determines the risk reward? Someone likely did a study and decided where to draw the line. Not much different than what Congress is asking the FAA to do.

How about back when all jets were three person crews? What suddenly made two person crews acceptable? I was not flying then, but I assume someone looked at the current/new technology and decided that it was acceptable for two people to do that. I imagine the flight engineers were complaining about that change, but these days I don't see anyone suggesting flight engineers be put back into the cockpit to increase safety.

What about ETOPS when jets moved from three/four engine designs over to twins? Who decided it was 'safe enough'? Probably a study. Four (equivalent) engines should be safer than two, so why were twin jets permitted?

Things change. In order for the FAA to gather data and decide what is and is not an acceptable risk they need to do a study. Just what Congress is wanting. If it is such a clear case of not being safe I would have thought being part of the study and leading it to a negative conclusion would be far more effective in the long term than simply delaying the study from being done.

[StudentPilot]

http://m.scmp.com/news/china/society/ar ... pit-window

FO injured / half sucked out, CPT loses all automation which presumably would have cut the ground to flight connection. This just happened and is another example of how 2 up front increases safety yet doesn’t show up in the statistics.

Perhaps, perhaps not. Maybe the ground link will be required to be in a hardened case in the tail so it is physically separate from the cockpit and independent of the flight deck automation as a result of the study Congress wants done.

One could also spin this the other way. If independent systems for the aircraft- and ground-based controls were required, one pilot gets half-sucked out the window and injured. Is it better to have a second pilot beside him with the wind and cold blasting into the cockpit, no automation, and potentially limited or lose of other controls or for a ground-based pilot sitting in a comfortable, climate controlled office with completely independent/redundant controls to take over?

[Rockie]

Do you have good data as to which one kills more people?

I can tell you how many times a week automation tries to kill me and I prevent it, but unfortunately those stats aren't collected by Boeing, Airbus or anybody else.

What you are flying is not an automated aircraft, and that is not what the discussion was about. We've discussed this before, and while I respect your skill and experience as a pilot, I also know you have little to no experience with system design or modern automation.

Well I'll have to disagree with you there. I do in fact fly an automated airplane and I have a lot, to a real lot of experience with modern automation in aircraft. I've also had the privilege of seeing how automation and systems design has progressed over the years. Some of it has made my life easier, some - like automation - has made it much more complicated and difficult to stay on top of. Because you see it cannot be trusted. It isn't a pilot and never will be until someone builds a real AI.

I admit I can be a dinosaur in many things, but I'm still very much smarter and capable - and necessary - than any automation I've seen.

[rookiepilot]

Because you see it cannot be trusted. It isn't a pilot and never will be until someone builds a real AI.
I admit I can be a dinosaur in many things, but I'm still very much smarter and capable - and necessary - than any automation I've seen.

Hmmmmm. Interesting comment.

Do you hand fly Cat 3 approaches because you don't trust the automation to fly the approach?

[Rockie]

Because you see it cannot be trusted. It isn't a pilot and never will be until someone builds a real AI.
I admit I can be a dinosaur in many things, but I'm still very much smarter and capable - and necessary - than any automation I've seen.

Hmmmmm. Interesting comment.

Do you hand fly Cat 3 approaches because you don't trust the automation to fly the approach?

I am on the controls all the way to a stop because a) it's required, and b) I don't trust the automation. Aircraft accidents happen too frequently because pilots "trusted" the automation when it didn't do what it was supposed to do, or wasn't designed to do it in the first place. You guys know all about that because you shit all over the crews when it happens. Are you changing your tune now?

CAT III requires an autoland because there is insufficient visibility to land or roll out visually. There are stringent weather, wind and airport equipment and operational limitations to meet before low visibility approaches are permitted, and we are trained and prepared to abort the landing as soon it starts to go wrong. If you do these approaches yourself do you slide your seat all the way back and tuck into a good book while it's going on?

[complexintentions]

@rookiepilot (Rockie beat me to the reply!)

A human pilot is still flying the aircraft on a Cat 3 approach. The land/go-around decision is made by a human. The automation simply holds the controls as dictated by the pilot. "Otto" is as dumb as they come.

You seemed to have missed the entire point that flying an aircraft is far more than manipulating the controls.

[rookiepilot]

Aircraft accidents happen too frequently because pilots "trusted" the automation when it didn't do what it was supposed to do, or wasn't designed to do it in the first place. You guys know all about that because you shit all over the crews when it happens. Are you changing your tune now?

Another interesting comment.

Isn't it also true aircraft accidents sometimes happen because pilots don't use the automation available, misuse or disable it, (like some annoying warnings disabled by flight crews that have contributed to past accidents) or misread the information provided by the automation?

As One example only -- the countless accidents that have occurred from simply running out of fuel, -- in spite of automated, distance rings available even on single engine aircraft PFD's -- prove the weak link in the chain, is far more often, the human one.

[complexintentions]

Now you're just arguing for its own sake.

If you're determined to place the blame on humans for automation-related accidents, you must also consider the humans that designed the man-machine interface and the automation itself.

Technology is not magic. It's imagined, built, and used by humans. All of the human errors are baked right in. Automation may perform certain singular mindless tasks very well, like pointing the aircraft in a instructed direction, but unless you have true AI (not algorithm, self-aware) it doesn't replace a human.

As rockie has stated repeatedly, it's just a tool to be used.

[C.W.E.]

UREKA!!!

I finally agree with Rockie.

T

[rookiepilot]

Now you're just arguing for its own sake.

If you're determined to place the blame on humans for automation-related accidents, you must also consider the humans that designed the man-machine interface and the automation itself.

Technology is not magic. It's imagined, built, and used by humans. All of the human errors are baked right in. Automation may perform certain singular mindless tasks very well, like pointing the aircraft in a instructed direction, but unless you have true AI (not algorithm, self-aware) it doesn't replace a human.

As rockie has stated repeatedly, it's just a tool to be used.

I do not disagree with any of this, or the need for human oversight.

The implication correctly programmed technology is more dangerous than human factors as a dominant cause of aircraft accidents, even in the ultra safe airline world today? Sorry, I'm not seeing that.

If there was such zero trust in the automation, as Rockie implies, required to fly a Cat 3 autoland, those approaches wouldn't happen, to such precise minima.

[complexintentions]

As One example only -- the countless accidents that have occurred from simply running out of fuel, -- in spite of automated, distance rings available even on single engine aircraft PFD's -- prove the weak link in the chain, is far more often, the human one.

Your argument is growing increasingly incoherent.

Your example mentions "automated distance rings" that didn't prevent a fuel starvation accident. But automation is the answer...?

Your statement about humans being the weak link in the chain is completely misleading. More correctly, it could be stated that in the absolutely overwhelming majority of flights the human factor is the reason flights are completed safely and successfully. And in a minuscule percentage of flights, a sequence of events - a "chain", if you will - of events led to an accident. Again, overwhelmingly, it has been shown over and over again that "human error" was only one factor in that chain. And ironically, in the modern age many of those errors are related to the design, use, misuse, malfunction, or over-reliance (aka TRUST) of...automation.

All of which does not in any support the premise that automation can ever eliminate human error as automation is conceived, built, and operated by humans.

It's Only A Tool.

As far as trusting automation on a Cat IIIb, there is so little trust in it as to require 3 redundant systems chock-loaded with warnings overseen by two humans to intervene the second there is a malfunction. LVO ops are so cumbersome they slow airport arrival rates to the degree that they're only used when visibility makes them absolutely necessary for landings. Meanwhile, the most efficient way for airports to sequence arrivals is using visual (aka human-flown, human-judged) approaches. Hmmm.

[Rockie]

Technology - correctly programmed or not - does not work in isolation to human factors Rookie. The human is the lynchpin, the Jesus nut if you will of the whole operation. It is up to engineers to build technology that seamlessly and reliably aids humans, not replace them. Because that won’t work in this business. Unfortunately pure engineers are the last people who should be entrusted with that task without direct human supervision. Kind of like automation when you think of it.

You also clearly have no idea of the concept or practical application of low visibility approaches. I suggest you drop that line of thought.

[rookiepilot]

Is airline safety many orders of magnitude greater today than in say, the 1960's, due to,
A) better training and understanding of human factors (CRM, Ect)
B) technology enhancements?

To be fair I would say it's a combination, no? ---

Anyway I've asked a couple of what I thought are reasonable questions, not really getting direct answers -- welcome to the internet -- because surely as the sun rises, greater automation is coming to a whole host of industries, whether we like it or not.

One day in our futures, likely long after I'm gone to be fair, the skies will be filled, Jetsons style, with thousands of automated aircraft. Battery technology, AI, and VTOL technology will mature enough to see this happen.

I have little doubt of this, nor the fact only computer guided aircraft could possibly maintain safe separation in such an environment. So Rockie, I think you're wrong, given enough time. Humans will be replaced, as they have in many other industries.

Look where autos are headed. It's coming, to a highway near you. Am I comfortable with this? Not particularly.

The computer, Deep Blue, defeating the leading chess grandmaster years ago, was a sign of things to come. I'm not sure I like the world that is being created, but it is what it is.

[Rockie]

True. Our safety record is a combination of human factor and technological advances.

Chess master beating computers are as Complex says, mathematical formulas, not self aware thought. Even in chess the variables are limited and given enough time possible to program. Flying, as is real life, is not so predictable.

[confusedalot]

wow.

And now for a reality check, 175 year old railroad technology (apart from the monorails you see from time to time) still require at least one operator, even with the far advanced bullet train system.

And that is on a huge piece of equipment that cannot go anywhere else than the rails it is on. Don't know how people can think that technology can do everything when even the most controllable of transportation modes have not yet achieved operatorless levels.

[Noo]

I've been on driverless trains.

[Rockie]

I've been on driverless trains.

Me too, including one that went 500 kph. But they were on elevated tracks with multiple safeguards that would bring the whole shebang to a stop at the first sign of trouble. If only we could do that at FL370 and 500 kts...

[goingnowherefast]

Don't forget that aircraft are much more reliable than they were in the 60s, and that makes them substantially safer too. How often were Super Connie's blowing engines vs 787s today.

The A320 is substantially more automated than the 737, yet both are exceedingly safe aircraft. Automation doesn't automatically make aircraft safer, it just provides alternate means to control the aircraft, often with increased capabilities and efficiencies. For example, CAT II and III approaches.