Brain vs computer

General Discussion
41

Does anyone here not fly because they don’t trust the computers which are in pervasive use during all aspects of a flight?

I worked on black boxes for most of 31 years until I retired. I will drive 3,000 miles rather than fly on a commercial airliner. The black boxes aren’t the only reason, but they are a reason.

As far as software, no one has come up with a perfect test yet. They keep working on it, but in the end complicated software can only be tested by long term usage.

Note that the partisan gridlock exists not because of bad legislators, but because the people themselves are sharply divided on those issues. To eliminate the gridlock can only happen by one side surrendering their moral views.

42

Trust the computers? It’s not the computers, it’s the lack of concern by the airlines, putting profits above passenger safety. I used to fly to NYC four or five times a year on long weekends. I just did it for fun, much enjoyed just to walk around the city. Since 9.11 though, and that I learned the airlines had prevented a requirement to strengthen the cockpit door due to the extra weight and therefore needing more gasoline to fly the plane, I haven’t returned. I only fly now when there is no other alternative, and for me there is almost always another alternative. My annual flying miles are probably down 95% compared to before 9.11.

43

"and that I learned the airlines had prevented a requirement to strengthen the cockpit door due to the extra weight "

Where’d you hear that? The doors have been strengthened, no question.

44

Yep and now the pilots can bolt the door from the inside and fly the plane into the ocean with no one being able to stop them.

45

"To eliminate gridlock can only happen by one side surrendering their moral views."
Gridlock is seldom about morality; it’s most often about economics. It’s about pacifying the donators to their reelection efforts and maintaining power and control. If it were about morality, we wouldn’t be killing people abroad and allowing them to die at home while poisoning our planet just to make the wealthy wealthier. Gridlock is a perfect example of amorality.

46
Isn't it only a matter of time before the technology for computers to operate an automobile proves it reliability and superiority?

Does anyone here not fly because they don’t trust the computers which are in pervasive use during all aspects of a flight?

Since the subject of flight has been raised a few times, let me chime in. As a certificated commercial pilot with 2 years’ line flying (prior to career-ending disability), I think I can contribute to this.

Automation has been a somewhat controversial addition to flying. Compaines love 'em, because they are extremely repeatable–they do the same thing the exact same way, every time, and never miss an item on a checklist. They generally write their company flight manuals to maximize the use of the autopilot…though I’ve heard they’re coming back around in philosophy.

The downsides of automation, as I see 'em:

  1. You lose flying skills. The story about senior pilots, on trans-Atlantic flights, needing sim time to keep their skills up is peculiar. The Asiana crash on landing was due to skill deficiency. To what extent did an lack of recent hand-flying contribute? Would the crash have happened if Asiana had a policy of hand flying approaches? Who knows?

  2. You give up some situational awareness of what’s happening to the plane. About 15 years ago, an ATR-72 augered in from a hold in known icing conditions. Basically, the plane iced up , behind the deicing boots, causing one wing to produce progressively less lift than the other. The autopilot just banked into the “good” wing, more and more, and the flight crew was unaware. When the condition exceeded the ability of the ailerons to correct, the AP kicked off…and handed the out-of-control airplane over to the crew.

At the time, much was made of needing to increase holding speeds, derate “known icing” certifications…I always noted that the accident would NEVER have happened if somebody was hand-flying. The FO would have noticed that he was having to bank into one wing, more and more, and they would have GTFO of the icing conditions.

2a) A L-1011 crashed, ages ago, because of a burnt-out landing light. The crew put it on AP, and all three of 'em tried to figure out if it was just the light or a bad landing gear. Somebody bumped the altitude hold off, and the plane entered a minutes-long, shallow descent into the everglades. Again, if the CA had somebody hand-fly, they’d be alive. (Heck, CA needed to designate SOMEBODY to “mind the store,” AP or not, while the other two troubleshot!)

3)the AP can and will kill you if you’re in a GIGO situation. I purchased a book that was a collection of cockpit voice recorder dialogue in a series of crashes. One very harrowing one was where a plane was returned to service, after painting…with the pitot-static system still taped off! Of course, Murphy had them take off into IMC.

They (kinda) ID’d a P/S system failure…and attempted to put it on AP while they ran the checklist! The computer was in a GIGO scenario, and–being a computer–had no plausibility check and tried to fly according to the worthless data, ending in the death of all aboard. The sad thing is that “pitch and power setting” the MOST BASIC thing to do in a multiple SHTF situ, would have saved everbody aboard! Even not knowing the equipment, I know they could have a) ID’d the problem, b) gone max continuous power + 5deg. nose up c) declare emergency d) get ATC vectors for e) an ILS approach. There are three, fully independent attitude indicators on a transport-cat aircraft…I’d trust my life (and everybody else’s, too) that at least one of them will ALWAYS work.

When I first tried for my IFR rating, my curmudgeon CFI MADE me memorize all the pitches and power settings for every phase of flight. But…your actions in an emergency are the ones you’ve committed to memory in practice. Obviously they were trained WRONG by a automation-loving company, which KILLED the crew, and all aboard. Yes, I feel that strongly about it, and I’d say that to the face of the author of the CFM.

So, now you know why I’m so skeptical about automation of the highways…I’ve seen the negative effects of too much automation in one transportation field, and I STRONGLY suspect similar effects would happen in a headlong rush to automation in another.

47

@meanjoe75fan. that was a good,informative post. however, you used many acronyms that were greek to me, so I only got about half of what you wrote

48

Let’s see…
“ATR-72” is a turborop commuter. “L-1011” is a tri-motor jet that hasn’t really been used since ETOPS (twin-jet flights allowed over the Atlantic).

“Deicing boots” are what they sound like: inflatable rubber boots on the leading edges of the wings. You inflate them to break off any ice that might be clinging to 'em.

“Known icing certification” means that the FAA says a plane is legal and safe to operate into clouds that have produced ice accumulation on other aircraft.

CA= Captain (pilot). FO=First officer (co-pilot).

GIGO is “garbage in, garbage out” and is a programmer’s term for getting worthless data from a program fed worthless inputs. (Often used in defense of a termed “computer error”: “it’s not the computer’s fault…GIGO…your data was worthless to begin with!”)

“Landing lights” are three green lights that illuminate when the gear is down and locked. There is also generally an “in transit” light, and some means of showing a fault. (A burnt-out light, BTW, is a very minor event…it happened to us, flying in to Atlantic City. We just told tower what was up, and that we’d need 5 minutes prior to landing to run some checklists. They said OK, but that they were going to scramble airport rescue regardless, on general principles.)

“altitude hold” is the portion of autopilot that holds (and captures) an altitude. Meaning, some of the AP can be on, without it all being on.

“pitot static system” are two sources of air pressure…the “static port” measures ambient pressure, and the “pitot tube” points directly forward so that it also measures the ram effect of the moving air. Together, they provide your data for airspeed, altitude, and rate of climb/descent. With them plugged, all of that data would be completely false!

“IMC” is “instrument meterological conditions,” which means you are in the clouds and cannot fly by reference to the outside horizon (you’d have no idea which way is up, which is what doomed that flight…visual flight would have given better references to replace the P/S stuff.)

“pitch and power setting” is where the nose is pointed, and how much power you’re giving her. If you “know your aircraft,” you should have a good idea what pitch she takes on climb out, at cruise, descent, and approach to landing…and where to set the thrust levers.

An “attitude indicator” is a gyroscopically-powered “artificial horizon” and is what you use in IMC to tell which way is up. They are probably visually familiar to the non-flying public…I’m sure Google links to a picture or two.

“IFR rating” is an additional endorsement that says you’re trained how to fly according to instrument flight rules, in IMC. The most basic license if VFR-only…meaning VMC.

“CFI” is certified flight instructor.

“CFM” is “Company Flight Manual” and is the system to which the company trains you and expects you to fly their equipment.

If I left anything out, somebody let me know…

Oh…and an “ILS” is a navigational beam that extends off the approach end of the runway, aligned with the runway (as much as possible) and on a 3-degree slope (as much as possible). In the airplane, there’s an instrument that looks like the crosshairs of a rifle scope…only the horizontal and vertical lines move. You correct towards the needle (i.e. fly to the right if the vertical needle is to the right of center)…so if the needles wind up centered like in a real rifle scope, you’re doing it just right.

49
It's not the computers, it's the lack of concern by the airlines, putting profits above passenger safety.

That is definitely a concern. All you have to do is take a look at all the security breaches businesses have had that effect you. Target is the biggest example. They KNEW there was a security problem. They didn’t have the right safeguards in place to prevent this from happening. In the past 10 years my account information has been compromised at least 7 times by companies I’ve done business with. Luckily I wasn’t in the Target breach. I never lost any money…but I had to get a new credit card. I have some automatic payments taken from my credit card (like Ez-Pass). To safeguard these systems - the companies will have to spend some money…in some cases a lot of money. And many are refusing to do it.

50

@meanjoe75fan‌:
Thank you for that great post. I agree, increased automation won’t have its drawbacks.

51

“That’s part of the reason. The other is programming methodologies. Using best practices help ensure that code meets the correct objective. By having good methods in place and making sure everyone ensures they are following those methods has far more of an impact on good software then a good test team and what methods they are using.”

100% accurate. There’s an old saying, you can’t test in quality. That has to be built in from the start. It starts with good product definition, hazard and risk assessment to identify where mitigations are required, coding standards and so on. The opposite approach is cowboy coding with back end testing. This is the microsoft approach and you all know how that turns out…

52
This is the microsoft approach and you all know how that turns out...

I’ve worked a lot with Microsoft development teams…and they have some of the BEST standards in the business. Their approach to software development is excellent. Maybe some groups don’t follow their own standards…but Microsoft standards are EXCELLENT.

53

But nothing will ever beat the human peripheral vision.
A computer can have dozens of peripheral cameras BUT to process all that movement and action plus analize any potential threat or not…AND keep driving in a smooth uninterupted manner…is human .
How much massive amount of peripheal data our minds process is so automatic and sub-concious that we don’t realize until we lose that ability.
Not untill my wife lost the vision in her right eye did we even begin to wonder tha amaizng machine of the human brain processing the sense of sight.

54

How many man hours would be needed to program a ‘robot’ to ride a bicycle?

55

Don’t underestimate the technology. Here’s info on Google’s cars:

And here’s a video of a near-blind ‘driver’ taking one to get a burger and his laundry:

56

Thanks, @texases, case closed.

57

@texases, yeah, that’s what I’m talking about (in the other thread on this topic).

When we talk about things like new technology, it reminds me of just what a group of curmudgeons we’re dealing with. Instead of evaluating how the technology actually works, you guys sit around imagining the worst case scenarios.

58

For those of us with ‘regular’ cars, I guess we could buy an Asimo to drive us around:

59

Are you kidding? Even when @meanjoe75fan turns 90, no computer will drive as well as he can! :wink:

60
Yep and now the pilots can bolt the door from the inside and fly the plane into the ocean with no one being able to stop them.
Bing

It has never been brought out in the news so I don’t know if this was investigated but it would have been possible for a passenger to fly that plane from with an iPad or similar wifi enabled device. Certain 777’s, of which that particular plane was one, used a shared network between the planes guidance system and the passengers wifi.

If the plane was in auto pilot, a knowledgeable passenger, one who had been an avionics technician and knew how to trouble shoot INS (inertial navigation system) problems by running a routine called a flycatcher would insert false heading information into the INS to see if it made the proper corrections. By doing this, you can make a plane fly in a different heading than the plane thinks it is flying, and the pilots would never know.

I guess this falls under the GIGO scenario.