2
$\begingroup$

I understand that TCAS can handle several "targets" simultaneously. However, I'm interested in the (highly hypothetical) scenario where TCAS would need to deal with multiple conflicts in a dense environment.

Is there a point where TCAS goes into some sort of "cannot resolve" mode becoming effectively inoperable, or does it try to cope with everything thrown at it, even if it leads to impossible solutions (defined here as maneuvers outside the flight envelope or the ability of crews to follow in a timely manner)?

$\endgroup$
  • $\begingroup$ Something like that has happened but TCAS wasn't really relevant. $\endgroup$ – Pondlife Oct 21 '18 at 6:29
  • $\begingroup$ You seem to be treating Air Traffic Control as a single, monolithic entity; it is not. ATC is the combination of a large number of facilities: everything from various area control centers right down to the tower controllers at a small airport who (especially by the time we factor in non-US airspace, which is a huge part of the world) may or may not even have radar service. $\endgroup$ – a CVn Oct 21 '18 at 6:45
  • 3
    $\begingroup$ So is your question more like "what does TCAS ultimately do if you keep throwing traffic conflicts at it?" than it is "can TCAS be overwhelmed by multiple conflicts?" ? If so, you may want to revise the title as well. $\endgroup$ – a CVn Oct 21 '18 at 17:20
  • 1
    $\begingroup$ I don't think it's a duplicate. This is a more specific question than the linked one. $\endgroup$ – Zeus Oct 22 '18 at 2:53
  • 1
    $\begingroup$ I've removed all references to ATC, since that is wrong - TCAS is not a replacement of ATC when the latter fails. In any case, the focus of this question is what will TCAS do if we keep throwing traffic at it until a resolution is impossible, which is interesting. The TCAS algorithm is deterministic and well-defined, so it is answerable. I think the algorithm is behind a paywall though. $\endgroup$ – kevin Oct 22 '18 at 8:59
4
$\begingroup$

From my general knowledge about TCAS (that is, not from actual simulations of such scenarios), I'd say that yes, it can be overwhelmed, but not in the sense that it will give up.

First, TCAS can demand a very limited range of resolution advisories (RAs). They are only vertical. The standard "Climb" or "Descend" RA assume the rate 1500-2000 fpm. TCAS can ask to increase it to up to 4400 fpm. (Or it can ask not to manoeuvre). In most installations, it will also not produce some RAs depending on the aircraft configuration; for example, it will not ask to climb at all if the airplane has flaps down past certain point.

Second, TCAS doesn't think too strategically. It can handle many targets ("intruders"), but in the end it will sort them according to their severity (primarily using the time-to-encounter factor) and will handle the most immediate ones in RAs.

It may well happen that after a short time, another intruder will become more severe. TCAS will then recalculate RA accordingly. (It takes into account the current climb rate and is generally reluctant to change the direction or even to choose a so called "crossing" RA, where the aircraft must cross the current path of the intruder, but it will do that if it determines this necessary. It also takes into account the typical (but clearly defined) pilot's reaction to the first and subsequent RAs).

So, at each point of time TCAS will "try its best" to handle immediate threats. It will (generally) not demand impossible manoeuvres nor superhuman abilities: all it can demand is known and accounted for. Yet, in theory I bet it is possible to create a scenario where TCAS would lead to very close encounters, or will select a sub-optimal strategy - even if we assume that all the aircraft obey their RAs precisely, which in itself is not given. In aviation, the engineers (so far) generally favour clear deterministic algorithms over something very clever but unpredictable and hard-to-test.

$\endgroup$
  • $\begingroup$ It seems that , based on the available information (short of receiving insights directly from the engineering team), this is the most complete answer we can have. I'll mark it as the accepted one, though of course if someone has more to add I would still be interested in more. $\endgroup$ – Digital Dracula Oct 22 '18 at 17:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.