News Germanwings Flight 4U9525 (Airbus A320) crash in Southern France (24 March 2015)

[DeskOrbinaut]

9-11 was only possible because of two things:

(a) passengers have been taught not to resist the hijackers

(b) shooting down a hijacked airliner was not considered an option

(a) access to the cockpit was not a big deal

(b) passengers have been taught not to resist the hijackers

(c) shooting down a hijacked airliner was not considered an option

Both of these circumstances no longer exist, so 9-11 style attack can no longer be repeated.

Until someone does. I have learned to use the word "not possible" with greatest care. Everything is possible. And someone else always will be "smarter" than you and me. Besides fighting of 4-6 hijackers can be a big deal.

Observe further that this is the second time reinforced cockpit doors have enabled hijacking by a crew member (third if you count MH370). The first was Ethiopian 702:

http://www.bbc.com/news/world-europe-26222674

So the question is why no lessons have been taken from Ethiopian 702.

http://news.aviation-safety.net/2013/12/22/list-of-aircraft-accidents-caused-by-pilot-suicide/

26 September 1976 - an Antonov 2 is not a commercial airliner so that's a little different story here

22 August 1979 - a stolen a military plane (by a mechanic) also is a different story

13 July 1994 - also no pilot suicide but hijacking a military plane by an engineer

21 August 1994 - one of the controversial "suggested" pilot suicide claims

Silk Air Flight 185 - controversial

11 October 1999 - the airplane was "stolen" on ground

31 October 1999 - might be pilot suicide

29 November 2013 - still no final report yet, happened in November 2013...

---------- Post added at 05:54 PM ---------- Previous post was at 05:35 PM ----------

Still, i find it hard to believe that someone can just sit there in front of the controls and breath normally while a mountain is getting bigger and bigger in front of you.

Not only this. It "was" (if...) a spontaneous decision. There are enough occasions during a normal working day of a pilot during which one pilot is alone in the cockpit because the other one is taking a leak. They flew from Dusseldorf to Barcelona as scheduled. But on the flight back he just decides to end it all right now within a few seconds. Does not seem reasonable to me. Or he was waiting for it for days/weeks/months already. But I simply can not believe this.

Also: where is the DFDR? The DFDR and CVR are both located next to each other in the tail of the airplane. Last report was that they "located" the DFDR. Now what?

The state prosecutor believes Andreas Lubitz did it on purpose just by being told from someone else what can be heard on the recording from the CVR 48 hours after the crash (actually only hours after they read the recording).

Sorry, too early for any certain conclusions. But maybe Brice Robin thinks more in favour of Airbus...

---------- Post added at 05:58 PM ---------- Previous post was at 05:54 PM ----------

A person with a heart attack is going to have a different breath pattern. This is why heart attack has been ruled out.

There is a bunch of other medical issues which can not be ruled out just by listening breath on the headset channel.

 

[ISProgram]

He was a pilot, he knew that soon after the crash SAR teams would've searched for the black boxes, he knew for sure how they work and the 30 minutes continous recording loop...I mean, you are desperate for some reason, you want to kill yourself, how could you do it without leaving a message?

Actually, that's a pretty good point IMO. If he was going to commit suicide, why didn't he disconnect the recorder (by tripping the circuit breaker)? That being said, not every incident of this type has had the recorders tripped.

 

[statickid]

<snip>

ACHTUNG!​

ALLES TURISTEN UND NONTEKNISCHEN LOOKENPEEPERS!​

DAS KOMPUTERMASCHINE ...</snip>​

Once I started learning german and realizing how similar to english it is (or vice versa, considering the evolution) this is a good example of how it started to look to me when I first started understanding it :lol:


regarding locked doors, freedom, and trust ... maybe we should just shackle the passengers with diapers and give the pilots a catheter and commode chair :hmm:

For hijackers, an overly complicated solution:
activate a remote control mode that can only be accessed when a tandem flying jet fighter ABOVE the plane relays encrypted instructions directly into a ground shielded radio port, while all on-board controls are completely disabled.

 

[RisingFury]

Ability for unconstrained navigation of the real world is one of the definitions of general artificial intelligence.
You'd have to get pretty close to that for planes to fly themselves with error rate similar to human pilots.

Navigation once in air is trivial. RADAR can provide collision avoidance with other airplanes. A global heightmap and RADAR combined can provide collision avoidance with terrain. Who says we need to limit flight computers to megabytes of memory?

The real issue is dealing with ATC and ground operations. Those aren't trivial. That's the part that really needs to be solved.

Computers as we know them suffer from GIGO problem - Garbage In, Garbage Out.
They don't make mistakes in their calculations, but they won't give you a correct answer from wrong figures.

But a human often can.
Because a human have a much greater map of the world than we currently can describe to a computer, and a pathfinding core that we don't really understand yet.
Summarily known as intelligence.

A human pilot can cross-reference multiple very noisy signal sources (eyes, ears, gauges n the plane, smell, "seat of your pants" and so on) and determine what is going on and how to get out of the situation.
A computer won't be able to solve an arbitrary sensor issue - if one gauge shows 10000 feet and the other shows 20000 feet, how would it know which is right?[/quote]

Some of the examples you people bring up against self-flying airplanes are just dumb...

Garbage like this is easy to detect. First off, don't include 2 sensors, include 4 or more. Second, make sure that different sensors work on different measurement methods. And third, it is possible to detect garbage data - compare them against expectations. If a sensor goes bad and start reading weird data, if it does so quickly, you know the plane can't change altitude very fast.

The big problem is not to define the algorithm of flying a plane - that is trivial.
The problem is in determining where the plane is in the grand scheme of things, what state is it in, what works, what doesn't, what is going on, where to go, and so on.

I don't think that's the issue either. Computer games have already demonstrated pretty sophisticated AI. Unreal Tournament 3 from years ago already has AI that learns from the players and exploits even stuff like map bugs to its advantage.

Autopilots are trivially simple programs. We're limiting them severely.

I don't think technology is advancing slowly. When I compare the flight deck of a 747-300 with a A320 and A380 it's an awesome progress.

The biggest change is the display of information.
There's also more flight envelope protection, but that's also trivial in comparison.

So, you have traded the risk of a pilot going rogue for a risk of a programmer going rogue, or worse: a minimum wage maintenance worker going rogue. And neither programmers nor maintenance workers undergo psychological evaluations...

A program can go under review and once it's decided that it works, it works. No more changes to be made.
A minimum wage maintenance worked could be prevented from tampering with the system by himself. You can make it so that multiple people have to be present to just open the access to the computer.

Care to throw any more preventable situations at me?




Ultimately the reason is fear. Nobody wants to fly in a self-flying plane. Nobody wants to make the first step. But they're coming as soon as computers can speak with humans and see where the taxiways and gates and terminals are...

As airplanes gain more sensors, it'll take more computing power and algorithms to integrate all that data.


Oh and by the way... the military is using self flying drones to kill people:

Reaper drone is capable of flying itself.

 

[ISProgram]

Wikipedia lists the flight as [ame="http://en.wikipedia.org/wiki/Suicide_by_pilot"]a case of pilot suicide[/ame] now, even though the investigation is still preliminary. :facepalm:

 

[Scav]

That's alright. The (social) media has already condemned the copilot anyway (at least everywhere I've seen so far). :dry:

 

[fort]

In other words, if the axe was used, we have to re-examine the situation based on possible reasons for the pilot to be in the passenger cabin with an axe...
Let's just say some of them would make locking the door to the cockpit look like a very reasonable thing to do

:yes:

That's a crow-bar, right?

There is another word in French: une pince-monseigneur.

 

[Artlav]

Navigation once in air is trivial.

Unconstrained navigation of the real world, not navigating in the air.
That means getting from one arbitrary situation to another.

Starting situation:
-Plane is high in the air over land with no fuel (too little fuel was loaded)
-Most of the sensors don't work, all that is available is GPS, some air speed indicators and a camera/radar.
-There are no known airports in glide range on the map

Target situation:
-Plane is on the ground with minimal or no casualties.

A computer that can solve this sort of problems would be G-AI or quite close.

And no, you can't just program every possible situation into it, since they are quite literally innumerable.

The real issue is dealing with ATC and ground operations. Those aren't trivial. That's the part that really needs to be solved.

THAT is the easy part. Just equip ATC with the right equipment to send common commands in digital form.
For uncommon commands see above.

And third, it is possible to detect garbage data - compare them against expectations.

That is the hard part. You as a human feel that it is easy, since your brain does such things automatically.
But try to make a computer do the same, and here comes trouble and yet-unsolved problems.

I don't think that's the issue either. Computer games have already demonstrated pretty sophisticated AI.

Ugh.
In a game the computer owns the world.

Think Orbiter - there are API calls that would give your precise position, speed, altitude above terrain and so on.
This makes writing autopilots almost trivial, since you're skipping right past all the hard parts.

I found that the hard way, when i built my quadrocopters.
There is no shortcut to make the flight computer aware of it's situation, you have to wrestle it out of the sensors.
Adding more sensors does not really help there.

A hover autopilot in Orbiter is a few lines of code anyone can understand.
A hover autopilot for a real world quadrotor is 1500 lines of code involving non-trivial math, and it still isn't something i'll let fly without having a kill switch in hand.

Reaper drone is capable of flying itself.

Hm?
I can't find any mentions of that, it appears to be a regular fly-by-radio drone.

 

[DeskOrbinaut]

The biggest change is the display of information.
There's also more flight envelope protection, but that's also trivial in comparison.

So the biggest change between black and white tube television and my i7 computer is the display of information...

Without offending you, but you don't really know much about the systems of airliners I think. Because the changes are immense. Not only in terms of displaying information. Early versions of the 707, 727 and 737 actually had no computer at all. The control of the engines was hydromechanical. Everything was entirely mechanical. And the "navigation computer" was nothing more than a plain calculator which just memorized and displayed coordinates. If there was a navigation computer at all. In the beginning there was nothing more than ADF and VOR navigation and radar vectoring.

Oh and by the way... the military is using self flying drones to kill people:

Reaper drone is capable of flying itself.

As I said earlier in the thread, a drone does not fly hundreds of human beings from A to B under various meteorological conditions in heavily used international airspaces. It takes pictures/recordings or kills people (wow nice, and even without lawsuit) mostly in restricted or prohibited airspaces above crisis regions in nice weather conditions. But still hundreds of crashes during the past 15 years.

 

[Urwumpe]

The hospital, in which Andreas L. was treated lately, has denied that he was hospitalized for depression.

 

[DeskOrbinaut]

The hospital, in which Andreas L. was treated lately, has denied that he was hospitalized for depression.

Thanks for that information.

I'm not really a big friend of the Springer press. But there is a good article regarding the premature conclusions by the French state prosecutor.

Sorry it's German only... http://www.welt.de/wirtschaft/artic...am-schnelle-Festlegung-auf-eine-Wahrheit.html

 

[Urwumpe]

Thanks for that information.

I'm not really a big friend of the Springer press. But there is a good article regarding the premature conclusions by the French state prosecutor.

Sorry it's German only... http://www.welt.de/wirtschaft/artic...am-schnelle-Festlegung-auf-eine-Wahrheit.html

Yeah, especially important in that article about what to expect: The french pilot trade union SNPL Alpa plans to sue against unknown for the release of trade secrets - they are also pretty upset about how much information is leaked from the investigation in a very short time and that already before the BAE had all pieces of evidence together and finished their investigations.

 

[Ripley]

...Also, you assume he would have someone to give a message to?

Well, his family at least...

 

[RGClark]

Deliberate crash highlights gaps in how pilots are screened.
BY CURTIS TATE AND LINDSAY WISE MCCLATCHY WASHINGTON BUREAU
03/27/2015 8:00 AM 03/27/2015 10:56 PM

Media outlets also reported that the torn up note declaring him “unfit to work” bore a code for depression. Under German law, the doctor would have been prohibited from presenting the note to Lubitz’s employer. Under Lufthansa regulations, Lubitz had no obligation to report the note to his supervisors.

http://www.miamiherald.com/news/nation-world/national/article16550852.html

This is surprising, if true. This would mean, if literally true, that for example if the pilot had a mental breakdown and was experiencing hallucinations then the doctors examining him could not report that to the airline. And the pilot would have no obligation to report that to Lufthansa. (Is the same policy true for other airlines?)

Would this also be the case with government agencies regulating airlines that the doctor could not report this to those agencies, nor would the pilot be required to?

If this really is true, presumably this would also be the case with physical conditions. So if the pilot was determined by a doctor to be susceptible to epileptic seizures that would cause him to blackout then the doctor could not report this to the airline, and the pilot would also have no obligation to report it(!)

Bob Clark

 

[DeskOrbinaut]

Deliberate crash highlights gaps in how pilots are screened.
BY CURTIS TATE AND LINDSAY WISE MCCLATCHY WASHINGTON BUREAU
03/27/2015 8:00 AM 03/27/2015 10:56 PM

http://www.miamiherald.com/news/nation-world/national/article16550852.html

This is surprising, if true. This would mean, if literally true, that for example if the pilot had a mental breakdown and was experiencing hallucinations then the doctors examining him could not report that to the airline. And the pilot would have no obligation to report that to Lufthansa. (Is the same policy true for other airlines?)

Just what I wrote on page 12: http://orbiter-forum.com/showthread.php?p=500867&postcount=180

The doctor does not even have to know your job. You can go to any doctor you want to. You are obligated to report more or less serious medical issues (or a stay in the hospital if it exceeds 12 hours or a medical certificate if it exceeds 21 days) to your aviation medical examiner which could then question your airworthiness. But if you don't... nobody will know. There is only one medical examination every 12 months by an aviation medical examiner and that's it. It's very easy to cheat. You can change your family doctor like your underpants... (and there is a pharmacy every 5 meters in Germany)

Actually I like the laws we have. It gives me freedom. I would call it one of the best healthcare systems. But maybe they should rethink something in case of aviation personnel...

 

[Urwumpe]

This is surprising, if true. This would mean, if literally true, that for example if the pilot had a mental breakdown and was experiencing hallucinations then the doctors examining him could not report that to the airline. And the pilot would have no obligation to report that to Lufthansa. (Is the same policy true for other airlines?)

Thats absolutely correct. And if you give your employer such an certificate, the copy for the employer does not contain your diagnosis. Only your health care insurance gets a single line ICD-10 summary. Before you complain, remember that usually this protects you as employee from being discriminated by your employer.

Also remember that the flight-medical certificate class 1 for pilots (and similar certificates in other similar risk jobs) is only valid for a year (if you are younger than 59 years. Later its only valid for 6 months) - pilot see their doctor way more often than normal employers, but these checks only care for physiological symptoms. And that with a very good reason: Without cooperation by the patient, no psychologist in the world can diagnose psychological diseases reliably. If somebody is intelligent enough to be pilot, he is also intelligent enough to "play sane".

And then: There are also many minor psychological diseases, for which you can be in treatment. Nothing that will affect your flightworthiness.

Of course you are now feeling badly vulnerable because of such an accident. But thats no reason to freak out. And stop posting his name - it is sick enough that the German media already forgets that the copilot also has parents and can't wait with drawing conclusions until the preliminary report is out.

 

[Ripley]

Just what I wrote on page 12...

Off topic: number of posts per page is a user-configurable option, so your linked post is on my 5th page.

 

[dman]

JAR-OPS 1.795 Crash axes and crowbars

(a) An operator shall not operate an aeroplane with a maximum certificated take-off mass exceeding 5,700 kg or having a maximum approved passenger seating configuration of more than 9 seats unless it is equipped with at least one crash axe or crowbar located on the flight deck. If the maximum approved passenger seating configuration is more than 200 an additional crash axe or crowbar must be carried and located in or near the most rearward galley area.

(b) Crash axes and crowbars located in the passenger compartment must not be visible to passengers.

Per wikipedia, the only A320-class a/c this would apply on is A321 with 1 class sit configuration (220 pax).

Crash Axe

http://www.allhandsfire.com/CRASH-AXES

 

[DeskOrbinaut]

A friend of mine will fly from Palma de Mallorca to Dusseldorf with a Germanwings A319 in 40 minutes from now (4U9585 scheduled for 17:00 CET). I'm curious if there will be special announcements from the cockpit/cabin crew...

---------- Post added at 03:23 PM ---------- Previous post was at 03:20 PM ----------

Off topic: number of posts per page is a user-configurable option, so your linked post is on my 5th page.

Thanks. It's post 180.

 

[SolarLiner]

Nice chart made from ADB-S data by flightradar24.com forum user solibra:

Original post