Andy Durbin wrote:
trajectory of each aircraft, but also predict collisions based on all
possible future trajectories for the next say 30 seconds. Try

Say 15 seconds, rather. I think it's useless to predict what will
happen in 30 seconds. Even if concentration is not at its maximum for a
moment, 15 seconds should be more than sufficient to avoid a collision
IF YOU SEE THE OTHER A/C.

resolving that mess when there are 30+ gliders at the top of the same
thermal waiting for a contest start. The false alarm rate would be
unacceptable.

I don't think so. I was at the ``kickoff-meeting'' of FLARM at the ETH
Zuerich, and I can assure you that the developers are very well aware of
this problem and have adressed it. There are algorithms which perform
well in that kind of situation and given that the workload is divided
among all 30+ FLARMs CPU performance is expected to be sufficient. The
developers cited extensive simulations which they did with IGC traces
and it seems that the ``false alarm'' rate was low.

That said, of course I can at the moment only cite and trust the
developers, and a system like FLARM needs to prove itself in practice.
There are too many factors which influence collision probability that
I'd dare a prediction here. But it's an interesting concept and if the
technical/regulatory question marks can be cleared up, we'll soon have
some data on whether it can help reducing accident rates or not.

Kind regards,
-Gerhard
--
Gerhard Wesp o o Tel.: +41 (0) 43 5347636
Bachtobelstrasse 56 | http://www.cosy.sbg.ac.at/~gwesp/
CH-8045 Zuerich \_/ See homepage for email address!