Kuro5hin.org: technology and culture, from the trenches
create account | help/FAQ | contact | links | search | IRC | site news
[ Everything | Diaries | Technology | Science | Culture | Politics | Media | News | Internet | Op-Ed | Fiction | Meta | MLP ]
We need your support: buy an ad | premium membership

[P]
Mid-air collision over Croatia

By thebrix in News
Wed Jul 03, 2002 at 01:17:22 PM EST
Tags: News (all tags)
News

I worked for many years on the development of the Swanwick air traffic control system, which manages en-route air traffic over the United Kingdom; thus the recent en-route collision over Germany is personally shocking.


It would be improper for me to speculate why it happened; many people have had their pound of flesh already. However, it is clear that en-route control is under suspicion. This type of control was introduced in full after the notorious Bryce Canyon collision of 1956, when it became clear aeroplanes were travelling too fast to be safely kept apart using visual aids alone, and that keeping apart ('maintaining separation') is its raison d'ętre. Clearly, in this instance, it failed; why it failed is for an investigation to decide.

However, I draw your attention to a little-known precedent; a collision between British and Yugoslav aeroplanes over Croatia in 1976. The circumstances are disturbingly similar - complex airspace (although not busy at the time), similar geometry of flight paths, not all staff present who should have been there, important equipment out of use - and I suspect that the current investigation will follow a similar course.

The analogy should not be taken too far - the Yugoslav controllers were using a now-antiquated technique called procedural control which relied on aircraft telling them where they are at intervals - but it is nevertheless striking.

Sponsors

Voxel dot net
o Managed Hosting
o VoxCAST Content Delivery
o Raw Infrastructure

Login

Related Links
o Swanwick air traffic control system
o en-route air traffic
o recent en-route collision over Germany
o had their pound of flesh
o Bryce Canyon collision of 1956
o maintainin g separation
o a collision
o complex airspace
o similar geometry of flight paths
o not all staff present who should have been there
o important equipment out of use
o procedural control
o Also by thebrix


Display: Sort:
Mid-air collision over Croatia | 60 comments (56 topical, 4 editorial, 1 hidden)
"busy airspace" (3.33 / 3) (#1)
by cockroach on Wed Jul 03, 2002 at 09:01:18 AM EST

AFAIK the airspace wasn't really busy at the time the crash in Germany happened...

However, interesting links.
--
Webisoder - never miss another TV episode

Workload (5.00 / 1) (#2)
by thebrix on Wed Jul 03, 2002 at 09:12:53 AM EST

It was night so I suspect that sectors were joined ('bandboxed') together; this is standard practice when traffic is low and allows fewer controllers to be used (and those that are being used not to become bored, that being dangerous in its own right).

Now, if the Russian media is to believed, there were only two Swiss controllers working that bandbox and one had left for a break. If that's correct there's big trouble afoot; I'm not 100 per cent up on United Kingdom procedures but I'm reasonably certain that, here, two people would be obliged to be working the bandbox at all times (probably by bandboxing yet more sectors, or bringing in spare members of staff).

Even with low traffic, one controller could become seriously overworked very quickly :/

[ Parent ]

Workload 2 (5.00 / 1) (#12)
by tftp on Wed Jul 03, 2002 at 09:58:31 AM EST

if the Russian media is to believed, there were only two Swiss controllers

It has been confirmed by Swiss already. One controller left for an "unauthorized" break, leaving only one at the ATC console. The one working could easily fall asleep without even noticing it (in the middle of the night, and with apparently nothing to keep him actively busy). That is, IMO, the most rational explanation of his actions.

[ Parent ]

Authorising a break (5.00 / 1) (#14)
by thebrix on Wed Jul 03, 2002 at 10:06:17 AM EST

Certainly, here, there would be a controller managing the shift and taking a general overview of what was going on.

The procedure would be that the controller who needed a break would message him or her (via a rough analogue of ICQ :) and they would then decide how to manage the break, as described above. Nipping out without doing so would (metaphorically) bring the ceiling down onto the controller who tried it.

(Food and drink is not allowed into the operations room for many reasons including the excretory one :)

[ Parent ]

Fixed (5.00 / 1) (#6)
by thebrix on Wed Jul 03, 2002 at 09:27:19 AM EST

Well spotted - I've replaced 'busy' with 'complex'. The Swiss accident occurred roughly where a number of flight information regions join, and the pertinent Yugoslav airspace consisted of three sectors one above the other. Vertical sectors are a necessary evil ...

[ Parent ]
I'm sure that damn US is behind this somehow (1.77 / 9) (#4)
by Sheepdot on Wed Jul 03, 2002 at 09:14:20 AM EST

That GD United States is to blame I'm sure. I can't wait till a BBC or Reuters article comes out implicating the US in a half-assed way so I can post a link and get +1 FP'd all over the place.

Got it (5.00 / 1) (#18)
by Herring on Wed Jul 03, 2002 at 11:15:03 AM EST

Malfunctioning TCAS (Traffic Collision Avoidance System) on the Boeing - once it detected the Russian plane descending, it should've changed to an RA Climb message.

Will this do?


Say lol what again motherfucker, say lol what again, I dare you, no I double dare you
[ Parent ]
That's it! (2.00 / 3) (#21)
by Sheepdot on Wed Jul 03, 2002 at 12:23:36 PM EST

You hit the nail on the head, son.

We all know that Boeing and Bush have been in cahoots since he was governor and guess what I found:

http://www.globalsecurity.org/org/news/2002/020205-dod02.htm

Also this gem about Russia:
http://english.pravda.ru/main/2002/05/24/29270.html

Boeing and Bush are working to kill suspected terrorist without the other governments knowing! More at 11!

[ Parent ]

Now there's lots of room in the air (4.50 / 4) (#5)
by sasquatchan on Wed Jul 03, 2002 at 09:25:48 AM EST

and the planes aren't that big relative to airspace in general. So to me, it is just unbelievable that two planes would manage to crash into one another. How could they have been setup on a crash course to begin with ? Then they both drop altitude to avoid ? What about swerving left/right ? I mean, sure you've got an additional dimension over cars, but you lack pesky curbs/trees/whatever that a swerving car will hit.. Yeah, it may take time to turn, but the TECAS should give plenty of warning, plus any in-cockpit radar.. I'm just so amazed that this could happen in 'this day and age'..

In closing, it must be a conspiracy! :)
-- The internet is not here for your personal therapy.

1:1E6 failure (4.50 / 4) (#8)
by wiredog on Wed Jul 03, 2002 at 09:41:10 AM EST

Actually, it sounds like a series of failures. The ATC may have messed up, apparently a safety system was turned off for routine maintenance. The procedures for collision avoidance may have contributed as well.

Pilots (I have a friend who's a commercial pilot) are trained to react immediately to certain warnings, and the responses are preset. When you get the collision warning you immediately descend. When you get the terrain warning, you immediately climb. These responses are preset so that the pilot doesn't have to take time, which he may not have, to think about what he should do. If you are 30 seconds from collision, and it takes 35 seconds to figure out the best way to avoid it, you're dead.

It takes less time to descend 100 meters, thus avoiding a collision, than it does to turn to avoid. It appears that both pilots took the same evasive manuevers at the same time. If just one of them had done so, or if they had done it at different times, the collision would have been avoided.

Can't sleep. The clowns will get me.
[ Parent ]

Or... (3.00 / 1) (#16)
by xriso on Wed Jul 03, 2002 at 11:08:08 AM EST

One guy was below the other, and the other dived.
--
*** Quits: xriso:#kuro5hin (Forever)
[ Parent ]
"Big sky, little plane" (4.00 / 5) (#9)
by BadDoggie on Wed Jul 03, 2002 at 09:44:42 AM EST

That's the motto that aviation used to use and small aviation still uses today. However, it just ain't so. You're moving more than 120mph/200kph in a machine that takes a relatively long time to convince that it wants to change direction. Furthermore, there are airways (just like highways), and most planes fly along these. Airways crisscross the entire planet between junction points. Vertical and horizontal separation is the key to keeping it safe, but sometimes it doesn't work. Only little itsy-bitsy planes like Cessnas and Diamonds and Archers can See and have enough time to still Avoid.

There is a chain of events that lead to an accident. Aviation is very focused on trying to always break any link possible. We have radar and all kinds of navigational aids including maps (VFR and IFR, with localised plates for every airport) radio aids (ADF, DME, VOR, LORAN, GPS) control towers and control centres and transponders and radar and Stormscopes and all kinds of stuff but shit still happens.

Happens a lot more on the ground, though.

Oh, I agree with another comment: this story needs to be pulled and resubmitted. The first paragraph or two (or just a teaser) go in the upper story submission text box and the remainder goes in the lower one.

woof.

Truth is stranger than fiction because fiction has to make sense.
[ Parent ]

Even faster (3.75 / 4) (#13)
by thebrix on Wed Jul 03, 2002 at 09:59:57 AM EST

The cruising speed of a 'standard' airliner is over 500mph (thus 1 mile travelled in roughly 7 seconds).

I heard that the DHL plane had its autopilot engaged at the time of the crash. An autopilot doesn't disengage instantaneously; add that to reaction times of controller + reaction times of pilots and a big distance may be covered before anything happens.

Someone told me that many Eastern European planes (not sure whether this one) have R/T transmissions broadcast through speakers in the cockpit, rather than through headphones. I don't like the sound of that (vital broadcast masked by clank of cutlery ...).

[ Parent ]

How fast? (4.00 / 1) (#41)
by tftp on Wed Jul 03, 2002 at 05:07:55 PM EST

An autopilot doesn't disengage instantaneously;

Actually, it does. On the airplane that I studied there are 3 ways to turn off the autopilot: by flipping a switch on the control panel, by pressing the override button on the yoke, and by applying larger than usual force to the yoke. All these methods instantly deactivate the autopilot; in 3rd method, in fact, you will be flying the airplane even before the autopilot gives up (it has stress sensors in control channels to measure the overload.)

It is also possible to steer the airplane while the autopilot is active. There is a 3D knob on the panel for that - very convenient, because you can make perfectly coordinated turns automatically.

many Eastern European planes have R/T transmissions broadcast through speakers in the cockpit, rather than through headphones

I doubt it. Firstly, the radio transceivers don't even have decent speakers, and they are meant to plug into pilot's headphone circuit. Secondly, there are two radio sets on Yak-42, and co-pilot can operate his set independently (and is supposed to). Speakers would make that impossible. Thirdly, how would one transmit without the headset - the microphone is mounted to the headset. I believe, what you heard is just a rumor.

[ Parent ]

Reality vs. what you heard (4.50 / 4) (#43)
by BadDoggie on Wed Jul 03, 2002 at 06:08:54 PM EST

The cruising speed of a 'standard' airliner is over 500mph (thus 1 mile travelled in roughly 7 seconds).
If you're anywhere close to the ground (i.e., "controlled airspace below 10,000 feet), your speed limit is 220 knots, unless Tower tells you otherwise (rare).

I heard that the DHL plane had its autopilot engaged at the time of the crash.
You probably also heard that TW800 was shot down because a lot of people swore they saw a missile. The NTSB has disproved it. You can believe the NTSB because they are 1) NOT political appointees and -more importantly and unfortunately -- 2) are not a regulatory body and have no enforcement ppower.

Autopilots take little effort to override, either with the flashy buttons or with the controls. I fly, and I've flown machines with autopilots. I had my instructors set me up with the worst situations possible so I could learn. One turned the autopilot on (small plane -- didn't shut off with different control movements) as I was on final approach and set it to 130 degrees to my left and 12,000 feet (in a Cessna 172!) with wing levelers on. I had to work the controls a bit harder until I caught that the damned thing was on, but I could've landed with the bastard still trying to pull me up. And that's in a little piece of private tin.

An autopilot doesn't disengage instantaneously; add that to reaction times of controller + reaction times of pilots and a big distance may be covered before anything happens.
More dreck out of the mouth of a non-pilot. Hit the off button and it's off instantaneously; move the controls opposite of what the autopilot wants and most shut off with a beep or three. This isn't meant as a personal insult, but you really have no idea what you're talking about. Could you please explain how an autopilot can affect a controller? (Hint: it can't.)

Someone told me that many Eastern European planes (not sure whether this one) have R/T transmissions broadcast through speakers in the cockpit, rather than through headphones.
"Someone" also told you you can make millions with bulk E-Mail advertising on the Interwebnet! All planes have speakers in the cockpit in case the headphones or their circuits die. There is no such thing in aviation as too much redundancy. In some small planes, people use and even prefer a hand mic and speaker; all planes have headphone jacks.

Concerning the mid-air over the Bodensee (Lake Constance), the reports I've seen show the Yak pilot failing to understand the order to descend, and by the time the pilot realised and reacted, Tower had already told the DHL flight to descend, putting them back on a collision course. 11:2 the fault is going to the Yak.

English is internationally acecpted and expected as the language for aviation. A Chinese pilot flying a Brazilian plane will talk to French controllers in English. However, the Austrian control center is sayin that there was no lanugage problem´and DHL is saying there was no problem with their plane, so we all get to wait on the black boxes. It still looks like the Yak was too late to respond, but I'll hold my judgement.

woof.

Truth is stranger than fiction because fiction has to make sense.
[ Parent ]

Various factors.... (5.00 / 1) (#45)
by phliar on Wed Jul 03, 2002 at 09:10:14 PM EST

Now there's lots of room in the air and the planes aren't that big relative to airspace in general.
You're not using the correct terminology. The approved way to say this is "Big Sky, Little Airplane."

Jokes aside, airplanes are not randomly distributed; they go to the same places. Additionally, the system of airways (or jet routes) we use makes them travel along a straight line between navigational beacons. Today, with area navigation (GPS, inertial etc.) being much more common, we should get away from this.

I'm not a controller; I'm a pilot. Furthermore I fly in the US, I've never flown in Europe. In my experience, it's more common for controllers to issue a turn for traffic avoidance rather than a climb or descent, since climbs and descents may be constrained by aircraft performance.

the TECAS should give plenty of warning, plus any in-cockpit radar
I assume you mean TCAS. By FAA regulations, cargo aircraft are not required to have TCAS. (So this may not have any bearing on the accident under discussion.) It's possible that the russian airliner did not have TCAS. It is also possible that there was no radar in the area, which may cause TCAS to not work. Additionally TCAS is far from perfect -- according to a controller I talked to a couple of years ago, they would issue false RAs quite often, and while following them the airplane would come too close to another, leading to a "deal" -- loss of separation.

In-cockpit radar is weather radar. Not only is it optimised for water droplets, it only looks straight ahead. Most mid-air collisions are not head-on.

Remember, none of this should be taken as any kind of theory on my part about the current accident. About that one I will only say that it's easy to put the blame on dead pilots.


Faster, faster, until the thrill of...
[ Parent ]

Lane Offsets (5.00 / 1) (#54)
by hughk on Thu Jul 04, 2002 at 07:42:29 AM EST

Ok lets drive along the highway together.

Lets all drive down the middle of the road.

Umm, did I say something stupid?

Regrettably, modern navaids are too accurate and aircraft do fly straight along the centre line of the air-routes and pass right over beacons. The proposed solution is that all aircraft fly slightly offset from the centre line, but that is only a suggestion. It isn't yet part of aviation law.

In this case the routes intersected on one level. Swerving is possible but relative velocities make the response more than a little tight. Swerving takes time, especially on a larger aircraft. Climbing takes even longer because of the time to spool up gas turbines. That is, you open the throttle and wait a good few seconds before the engines do anything. The only fast manouvre is the dive.

The only onboard radar specifically designed for spotting other aircraft and ascertaining their intentions is the TCAS unit. The indicator is common for the weather radar, so it isn't like that you ignore it. The problem is that until you get a transponder return, you only know that there is something near you, not their altitude. Aircraft pass over and under each other all the time, so you don't take that much interest until you get their altitude and heading.

[ Parent ]

'Non-central flying' (3.00 / 1) (#55)
by thebrix on Thu Jul 04, 2002 at 10:06:38 AM EST

This is already done in the United Kingdom; a flight can deviate up to 3 nautical miles from the centre line before it becomes 'off-route' (which automatically flags it to ATC) ...



[ Parent ]

Climb not necessarily slow (4.00 / 1) (#57)
by phliar on Fri Jul 05, 2002 at 05:30:40 PM EST

Climbing takes even longer because of the time to spool up gas turbines. That is, you open the throttle and wait a good few seconds before the engines do anything.
This is true for a sustained climb -- if you are told to climb to a different altitude, you need to open the throttle. For evasive action, it's different. At jet speeds (around mach number 0.8) a pitch change of a degree will be enough to give you a climb rate in the range of thousands of feet per minute. Of course your airspeed will decay but you don't care about that.


Faster, faster, until the thrill of...
[ Parent ]

ATC Failure? (4.50 / 2) (#7)
by wiredog on Wed Jul 03, 2002 at 09:34:20 AM EST

This CNN story would seem to indicate that.

Can't sleep. The clowns will get me.
Some oddities of timing (5.00 / 2) (#10)
by thebrix on Wed Jul 03, 2002 at 09:45:42 AM EST

As flight plans are filed before taking off and modern radar and flight data systems identify and track the planes from a long way back there should've been easily 15 or 20 minutes' warning to Swiss ATC who, in any case, probably would only be controlling the planes for 4 or 5 minutes anyway before handing them over.

I suspect that the gap between these numbers, and the 50 seconds or 2 minutes warning various sources are saying was given to the Russian pilot (what was happening in the middle?) will be of great interest to the investigators.

Personally, I would be very interested in the automatic warning the Swiss controllers should have been given but were not because the appropriate system was supposedly switched off. If it's what I think it is I'm amazed it was switched off without backup; there are systems called STCA (short term conflict alert) or, sometimes, MTCA (medium ...) which thrash through various algorithms and flash up on the controller's screen when two aeroplanes are deemed to be in danger of getting too close together; STCA, in particular, is absolutely fundamental.

[ Parent ]

Oddities (4.50 / 2) (#15)
by tftp on Wed Jul 03, 2002 at 10:06:55 AM EST

the 50 seconds or 2 minutes warning various sources are saying was given to the Russian pilot (what was happening in the middle?)

Nothing; those 2 minutes claim has been retracted by the Swiss already, and Germans now insist that the warning was given at -50 seconds, and the confirmation came at -25 seconds. It still doesn't say anything about the actual evolutions of the airplane(s) because pilots are trained to act first ("fly the plane!") and chat with the tower later. So the pilots could easily start the descent before they confirmed the order. This will be determined as the tapes of ATC and FDRs are processed.

[ Parent ]

Wrong 'middle' (my mistake) (none / 0) (#30)
by thebrix on Wed Jul 03, 2002 at 12:54:11 PM EST

Apologies for being unclear: the interesting bit is what happened between the warning of the flight and the first intimation to the aircraft that there was something wrong (a gap of 13 to 18 minutes, it would seem), during which the flight was almost certainly accepted into Swiss airspace and the beginning of coordinating it into German airspace probably took place.

I have a more than sneaking suspicion that the workload of the solo controller (low traffic volume, charge of a relatively large area, at least one conflict) is going to play a big role in all this.

[ Parent ]

Suspicions (none / 0) (#40)
by tftp on Wed Jul 03, 2002 at 04:50:46 PM EST

I have a more than sneaking suspicion that the workload of the solo controller (low traffic volume, charge of a relatively large area, at least one conflict) is going to play a big role in all this.

Yes, it is pretty much guaranteed. Even if pilots did not act at their best (at night, and with conflicting advisories from the controller and the radar, and with only seconds to react), the controller is the obvious primary cause of the disaster. You can't plow a car into a group of people and then blame the victims for being too slow to run away. You shouldn't be there to begin with; the Swiss controller had no reason to tempt fate playing with mere seconds.

Pilots could not see each other, and they (as reported) were not advised by the controller about each other. The controller was the only person - until seconds before the collision - who knew what plane is where, and he didn't tell anyone. For that alone he will be definitely reprimanded, if not indicted. And the other ATCO (who left his post without permission) is in it too.

[ Parent ]

Why not (4.50 / 2) (#11)
by Herring on Wed Jul 03, 2002 at 09:52:34 AM EST

See what people who know about aviation are saying?

Admittedly, I shouldn't post this as at time like this, their server has a hard time. Oh well.

The verdict at the moment seems to be that both the DHL plane's TCAS and the Swiss ATC made the wrong decision at the wrong time.


Say lol what again motherfucker, say lol what again, I dare you, no I double dare you
[ Parent ]
When you get a collision warning -> descend! (4.33 / 3) (#19)
by poopi on Wed Jul 03, 2002 at 11:33:35 AM EST

OK I've heard that on TV. I've read that in many sources - it still DOES NOT MAKE SENSE!

By definition two planes approaching each other on a collision course would BOTH get a "collision warning", BOTH as a rule would descend - which, unless someone can correct me, STILL keeps them on a collision path!!!

Would it not make more sense to turn the plane to one side (left or right) since that would result in a non-collision path? Sorry, for the YELLING. It just doesn't make any sense to me and nothing bugs me more than (apparent, in this case, [due to my ignorance, probably] stupidity).

-----

"It's always nice to see USA set the edgy standards. First for freedom, then for the police state." - chimera

Well, I don't think that is correct (5.00 / 1) (#22)
by theshunt on Wed Jul 03, 2002 at 12:32:56 PM EST

If two airplanes are coming toward each other head-on, I believe both planes are supposed to make a right-hand turn. However, this is the standard Canadian procedure, and it may be different in other countries. However, having both aeroplanes descend is quite a stupid plan.

[ Parent ]
Turning is relatively slow (5.00 / 1) (#28)
by thebrix on Wed Jul 03, 2002 at 12:48:30 PM EST

I'm surprised at this (as far as I know UK procedure is always ascend/descend) but am willing to be proved wrong ...

[ Parent ]
Expedite (5.00 / 1) (#42)
by I am Jack's username on Wed Jul 03, 2002 at 05:34:49 PM EST

I was told by a Cessna (small, slow plane) pilot that the international (for obvious reasons) rule to avoid contact with others, is for both parties to turn right. I think this is standard for planes, skydivers, kites...

This works fine in averting an otherwise head-on collision, but is can be troublesome at angles where one party would then fly towards another slower reacting plane.
--
Inoshiro for president!
"War does not determine who is right - only who is left." - Bertrand Russell
[ Parent ]

Standard in the U.S. (5.00 / 1) (#44)
by epepke on Wed Jul 03, 2002 at 08:10:15 PM EST

That's the standard in the U.S. as well, for general aviation. I suspect it's world-wide.

However, it doesn't work so hot at 1200 MPH relative. Planes have these wing thingies that make them wider than they are tall, and banking for a curve makes them taller as well.


The truth may be out there, but lies are inside your head.--Terry Pratchett


[ Parent ]
Not just Canadian but ICAO (4.50 / 2) (#46)
by phliar on Wed Jul 03, 2002 at 09:31:01 PM EST

Yes, the regs say both turn right. However, that is only for the head-on situation. And even more, that sort of thing can only happen when the pilot sees the other aircraft. An airliner at FL350 would just follow ATC instructions, and any TCAS warnings. You can't really expect "see and avoid" to work at those sorts of speeds.

For the controller, the appropriate action would depend on their established procedures and would probably depend on airspace considerations, other traffic, etc.


Faster, faster, until the thrill of...
[ Parent ]

Onboard collision warnings (TCAS) (5.00 / 1) (#27)
by thebrix on Wed Jul 03, 2002 at 12:47:13 PM EST

If both conflicting flights are equipped with TCAS, I believe it orders one pilot to climb and one to descend; if only one is equipped the order is always to descend (so ATC would tell the other to ascend). Obviously something went terribly wrong in this instance.

En passant, the fact that it takes a long time for the 'entire world fleet' to become equipped with a technology is a problem; TCAS is only about 10 years old, a mere youngster in aviation terms.

[ Parent ]

TCAS talks but there are s/w rev level issues (5.00 / 1) (#39)
by hughk on Wed Jul 03, 2002 at 04:46:33 PM EST

TCAS is a neat little system, but when you get the "RA" alert, you are instructed to take immediate avoiding action. You also alert ATC and surrounding aircraft with the message "TACAS Descend FL360". This has immediate precedence over anything else including ATC instructions.

The other TCAS equiped aircraft should receive the counter directive, however there are issues when one has started a TCAS descent and the other has not registered it yet. There should be a renegotiation, but in my understanding this only works with the very latest s/w releases.

I should add that this comes from The Professional Pilots Rumour Network. I have never flown a plane with TCAS (or anything larger than a Piper or Cessna), so I can't comment on the veracity.

The system simply an aid. Eyeballs take precedence, whether on the ground looking at the RADAR or in the air. However it appears that only one set of eyes was looking on the ground and in the air, it was dark and the paths intersected at right angles.

[ Parent ]

Collision avoidance (5.00 / 1) (#48)
by phliar on Wed Jul 03, 2002 at 09:46:53 PM EST

By definition two planes approaching each other on a collision course would BOTH get a "collision warning", BOTH as a rule would descend - which, unless someone can correct me, STILL keeps them on a collision path!!!
As usual, the media just misreports a hash of a bunch of different things. For a TCAS RA (TCAS is a transponder-based collision avoidance system; and RA is a resolution advisory, a command that the TCAS unit gives to the pilot) the idea is that when you have two airplanes, the TCASs will negotiate and one would climb, the other would descend.

In the "good-old" days (pre-TCAS, in visual conditions) aircraft on a head-on collision course would each turn to the right. Aircraft under ATC control would be kept separated by ATC. (At jet speeds "see and avoid" doesn't work too well.)


Faster, faster, until the thrill of...
[ Parent ]

TCAS? (3.00 / 1) (#20)
by wji on Wed Jul 03, 2002 at 12:19:27 PM EST

Maybe Russian airplanes don't have it, but I know Transponder Collision Avoidance System is designed to protect this. Do you have any idea if Aeroflot planes do have TCAS, and if so wouldn't it properly prevent this?

In conclusion, the Powerpuff Girls are a reactionary, pseudo-feminist enterprise.
TCAS (4.50 / 2) (#23)
by myshka on Wed Jul 03, 2002 at 12:33:21 PM EST

Yes, the Tu-154 had it. Whether it was working or not is a different issue altogether, since while TCAS is mandatory on European flights, operators have a 7 day window to operate an aircraft with a broken TCAS before fixing it.

wouldn't it properly prevent this?

It should. But then again, we all know that neither machines nor humans are infallible.

[ Parent ]

Hey myshka (1.40 / 10) (#26)
by Sheepdot on Wed Jul 03, 2002 at 12:45:30 PM EST

Sorry my comments about the US government conspiracy offend you(http://www.kuro5hin.org/comments/2002/7/3/85336/86889?pid=18#21).

Perhaps next time I'll have a link to Indymedia in there so you'll know its from a credible source.

[ Parent ]

Hey sheepdot (2.33 / 6) (#29)
by myshka on Wed Jul 03, 2002 at 12:51:31 PM EST

Your being a moron in a story that deals with real people's deaths and serious safety issues is what offends me.

[ Parent ]
really? (1.00 / 7) (#31)
by Sheepdot on Wed Jul 03, 2002 at 01:01:24 PM EST

Cause I really enjoy the fact that you just rated a comment down that you also responded to. You're a quality K5 member.

[ Parent ]
You're the one (3.00 / 8) (#32)
by wji on Wed Jul 03, 2002 at 01:06:21 PM EST

Who posts totally fucking irrelevant crap that wouldn't even be a good comment if it was in the right place. Zero!

In conclusion, the Powerpuff Girls are a reactionary, pseudo-feminist enterprise.
[ Parent ]
You're an idiot (2.87 / 8) (#34)
by dipierro on Wed Jul 03, 2002 at 01:09:50 PM EST

And I bet you +5 me for saying that...

[ Parent ]
Sorry... (2.75 / 4) (#36)
by zonker on Wed Jul 03, 2002 at 02:14:01 PM EST

but I've seen one or two other people mention this - where in the k5 guidelines does it say it's bad to rate and reply to a comment? I don't see a conflict of interest here. If someone can point me to a place in the FAQ where it says this is a Bad Thing, I'd appreciate it.
I will not get very far with this attitude.
[ Parent ]
it's not, just common courtesy (1.00 / 2) (#50)
by infinitera on Wed Jul 03, 2002 at 10:29:24 PM EST

Most people consider rating a form of replying - so, by both rating and replying, you are getting twice the voice of others.

[ Parent ]
Rated down. (3.00 / 4) (#51)
by mrgoat on Wed Jul 03, 2002 at 10:39:24 PM EST

I think it's more that someone would rate a comment down, that they apparently thought was worth replying to. If it's good enough to reply to, why not at least give it a 3?

"I'm having sex right now?" - Joh3n
--Top Hat--
[ Parent ]

agreed, check my user info;) [nt] (2.66 / 3) (#52)
by infinitera on Wed Jul 03, 2002 at 10:40:35 PM EST



[ Parent ]
Interesting take... (3.00 / 2) (#53)
by zonker on Thu Jul 04, 2002 at 03:40:26 AM EST

but I would prefer it if people who rate my comments also give some commentary. If I get a 1, I'd like to know if it's just because the person doesn't like what I've said or whether they think I said it badly , or both, or if they just don't like comments posted by people whose nicks start with "z" -- and I'd also like to know why people give higher ratings. Wouldn't it also be common courtesy to let someone know why they're getting the rating that they're getting? Actually - if I had to choose one or the other, I'd pick a reply to any comment I made. That's one of the reasons I visit k5, after all - for the occasional intelligent debate or discussion. (Rare, but it does happen.)
I will not get very far with this attitude.
[ Parent ]
attn: Type-R and rdskutter (none / 0) (#56)
by infinitera on Thu Jul 04, 2002 at 11:34:57 AM EST

Look, we tried to answer an honest question. We had a normal discussion. Offtopic does not exist here, for one, and two, our posts were still useful in some way to zonker (I assume, since he responded). If the discussion doesn't interest you, just ignore it. Don't rate it one and pretend it isn't valid discussion. Thanks.

[ Parent ]
'Should' to both questions (4.00 / 1) (#33)
by thebrix on Wed Jul 03, 2002 at 01:07:49 PM EST

TCAS should've been installed in all Russian planes entering European airspace by early 2001. I can't find any followups saying whether it was implemented or not.

All these technological aids, whether on aeroplanes or on the ground, are an immense aid to understanding but, in the end, human beings fallibly interpret their results. I think the uncrashable aeroplane or the remotely-piloted one are probably pipe dreams; the behaviour of metal tubes in a turbulent atmosphere can be tricky to predict.

[ Parent ]

What's the TCAS algorithm? (4.50 / 2) (#37)
by GGardner on Wed Jul 03, 2002 at 03:07:27 PM EST

We've heard a lot about the TCAS system, which operates on each plane, and tells the pilot to immediately ascend or descend to avoid a collison.

I can't find anything on the net about how it works though, and none of the standard news sources give you this much detail -- seems like a tricky algorithm -- you'd want it to work every time, and you'd want it simple enough so that many people could inspect and verify the implementation of the algorithm.

The obvious algorithm, "the plane which is already higher climbs, and the lower one dives" requires both planes to know each other's altitude precisely -- do they know this?

I suspect it's proprietary (5.00 / 2) (#38)
by thebrix on Wed Jul 03, 2002 at 04:37:56 PM EST

I've seen algorithms for the ground-based analogues (STCA and MTCA) and they are both proprietary and enormously complex; the one I saw had 40+ pages of vector calculus, fluid dynamics and whatever else you wanted to throw in.

This is because conflict algorithms try to solve a difficult problem; you have two, or more, aeroplanes each of which can be anywhere in a three-dimensional 'envelope' a given time later. In practice each plane would not be far from the maximum theoretical extent of that envelope; among other things, the non-uniform atmosphere introduces error. The whole issue of conflict detection, profile prediction and so on is a red-hot topic in algorithms (see a very nice summary in Acrobat format).

In ATC at the moment planes are told to fly at a certain level, rounded to the nearest 1,000 feet (for example, FL350 = 35,000 feet). Obviously there is some deviation from this in all three dimensions (due to, among other things, inaccuracy of instrumentation, particularly altimeters) so it is never known exactly at what height the planes are flying. Hence a great irony; if no action had been taken by anyone the German collision may not have happened. (Another useful reference on reduced vertical separation minima; I can tell you that the decision to space flights 1,000, rather than 2,000 feet, apart was taken after huge deliberation).

Eventually GPS-related solutions will appear which will give precise positioning. What that means for ATC is to be determined, but here's a taster :)

[ Parent ]

TCAS (3.00 / 1) (#49)
by ReverendX on Wed Jul 03, 2002 at 09:48:08 PM EST

Actually, if both planes have TCAS, it works fine. However, in this case, I believe only one of the planes had TCAS.

Being able to piss in an allyway is however, a very poor substitute for a warm bed and a hot cup of super-premium coffee. - homelessweek.com
[ Parent ]

now-antiquated technique called procedural control (3.00 / 1) (#47)
by ReverendX on Wed Jul 03, 2002 at 09:46:26 PM EST

Actually, there are many areas of the USA, and I presume the rest of the world that are not covered by radar. (A main area being over the oceans!) In these areas, the air traffic controllers rely on position reports. It's actually quite safe, however it reduces the amount of traffic that can be allowed in an area.

Being able to piss in an allyway is however, a very poor substitute for a warm bed and a hot cup of super-premium coffee. - homelessweek.com

Random flightpaths (5.00 / 2) (#58)
by kaet on Sat Jul 06, 2002 at 08:36:47 AM EST

A very small proportion of much of the flyable atmosphere is occupied by aeroplanes. There are places where this is not true, for example near airports. But these places are few and far between (though all journies fly through them at least twice!).

For the rest of space, the probability of intersection of two volumes the size of an aircraft in the volume of space is really rather small. Pick a bit of atmosphere at random, what's the probability it will contain an aeroplane?

In what I'll call free-air (air not near airports)  this is skewed by desired entry and exit points and the shortest path between them. Given free reign, all planes would travel by the most fuel or time efficient route between the two areas of non-free air.

Air traffic control forces airlines to act suboptimally, rising to a particular altitude, travelling along a particular non-minimal route, and so on, a cost in terms of efficeincy to ensure safety and to avoid what you might call the tragedy of the calculus of variations.
However, the algorithmic means by which air traffic control operates, which is predicated on the discretisation of space -- the division of space into discrete quantities, levels, paths, and so on -- combined with the increasingly accurate following of these paths by aeroplanes equiped with sophisticated navigational aids, is an initial step towards greatly increasing the probability of collision. Two appropriately timed planes travelling at 12,000 feet will crash: two planes travelling at a (truely) randomly chosen height between 11,000 and 12,000 will only have a 1% chance of colliding.

Of course, the algorithm layered on top of this discretisation, whatever it might be, is some sophisticated avoidance algorithm, or other. What is implemented, however, is not the algorithm, but a realistion of the algorithm -- a shadow of the pure algorithm cast on the wall of the cave -- with a certain probability of failure due to inevitable imperfections (human or otherwise). When the implementation fails, the accurate following of the discretised space increases the chance of collision greatly.

In freeair, it seems to me, there is a case for random flightpaths. These would be designed to alter the deviation from the mean path at the socially acceptable compromise between safety and price (which there is at the moment and, as I said above, in my opinion inevitable in any realisation, though discussion of this subject is taboo).

Note that I'm not advocating a free-for-all. That would lead to the tragedy of the shortest path again. There might be a box, sealed and sacrosanct, which emits a truely random path which must be followed in all its peversity.
Dispite what some of my more actuarial friends might claim, though, there is (unfortunatley for my safety) more to risk than the sum of probability times severity. It can be seen more clearly in transport accidents than in any other field, that our society firstly requires a moral agent to be given a thing we call "control" of a dangerous device (though that control is always imperfect and transitory, and secondly that a failure through imperfect realisation is actually the preferred form of failure, compared to failure through "random fate". It seems to me that this is largely through the blame structure we use to dispell accidents, which is well-defined in the case of perfect realisation, but is silent for "cold" "unfeeling" random error. An accident caused by random chance would be more difficult to dispell than one caused by imperfect realisation.

Given these societal preferences, it is perhaps best to approach a compromise system which fulfills the requirement of an algorithm which, when perfectly realised is completely safe, but within this constraint attempts to handle errors in realisation through randomising within the constraints which this imposes. For example a number of corridors containing a truely massive bundle of  randmly chosen (sometimes intersecting) paths. There /should/ be just one plane in each bundle, but when there is not, then they are likely to be travelling along different paths, and the probability of intersection is small.

Sorry this comment is so long.


This is called 'free flight' (none / 0) (#59)
by thebrix on Sat Jul 06, 2002 at 12:15:35 PM EST

You've described the Holy Grail of air traffic management; a Google search on 'free flight' will turn up loads of references.

Why it isn't done yet is because:

i. the technology isn't quite there to get precise positions of flights at all times. You can buy a GPS unit from a shop, but the requirements for big planes travelling fast are more exacting; 'oh dear, it doesn't give a reading' once is once too many;

ii. predicting conflicts is difficult and gets harder and harder the further back in time you go before the predicted conflict;

iii. with current flight volumes and degree of automation controllers have to channel flights along predefined routes because their workload becomes too great otherwise.

As you can see, i and ii together break the constraints in iii (to cut a long story short) and free flight, in effect, gives aeroplanes more autonomy and makes ATC more strategic (long-term) than tactical (short-term).

Obviously this is a good thing, but there's a huge organisational effort needed and massive re-equipage (mixing free-flight-enabled and non-free-flight-enabled traffic will be an interesting challenge ...).

[ Parent ]

Thank you (none / 0) (#60)
by kaet on Sat Jul 06, 2002 at 12:53:50 PM EST

Thank you for the pointer, it was interesting, but I don't htink this is quite what I'm describing (though it is very close, and much better than the present situation). I worry about the same three points for free flight. In particular, it allows /arbitrary/ positioning of aircraft rather than /random/ positioning, which is much worse in terms of risk of collision. Without any planning, there's also the risk that an aeroplane will be 'boxed in' in alert zones, particularly if you allow pilots to plot their own routes towards the edge of the freeflight zones (rather than to insist on them being random). As for i, knowing the precise position of flights at all times isn't important in a random system. An aeroplane might deviating from its course, in a random system, is not significant in terms of impact probability unless the error is systematic and non-monotonic across planes. As for ii, in a random system you don't predict a conflict, you design so that it is unlikely. (In composite systems you combine this with ATC). iii is true, but moving within a discrete set of paths within a channel (in a composite system) can be automated and not add to the ATC workload.

[ Parent ]
Mid-air collision over Croatia | 60 comments (56 topical, 4 editorial, 1 hidden)
Display: Sort:

kuro5hin.org

[XML]
All trademarks and copyrights on this page are owned by their respective companies. The Rest © 2000 - Present Kuro5hin.org Inc.
See our legalese page for copyright policies. Please also read our Privacy Policy.
Kuro5hin.org is powered by Free Software, including Apache, Perl, and Linux, The Scoop Engine that runs this site is freely available, under the terms of the GPL.
Need some help? Email help@kuro5hin.org.
My heart's the long stairs.

Powered by Scoop create account | help/FAQ | mission | links | search | IRC | YOU choose the stories!