Hacker News Reader

Comment by anamexis | Hacker News Reader

GPS surely is not used for instrument landings, and especially not for altitude.

>> GPS surely is not used for instrument landings, and especially not for altitude.

It is. And increasingly so in fact, as it allows more airports that don't have the resources to install and maintain ground-based equipment to have previously unavailable approaches. https://www.flyingmag.com/everything-you-need-to-know-about-...

However, it doesn't fall into exactly the same class as ILS approaches.

Instrument rated pilot here. A GPS LPV approach is essentially the same as an ILS [edit for clarification: a category 1 ILS]. It’s true that an LPV approach is classified as non precision but in practice an LPV approach is pretty much indistinguishable from a cat 1 ILS, which is classified as a precision approach.

Hm, that's interesting to me. Is it just regulators (FAA I guess?) falling behind the times? Or is the GPS version really not adequate for some use case that is only satisfied by ILS (not sure what that is but I assume it's the "gold standard" landing approach?)

There are different categories of ILS approaches (one of them is actually impossible to implement at this time, as it requires tech that is at best in testing), and most importantly GPS is not considered "safe enough". Most use in heavy aviation is due to easier operation (especially with autopilot) but in better conditions only (i.e. at best equivalent of ILS Category 1)

GPS notoriously has poor altitude resolution. For this reason, GPS approaches are technically not considered "precision" approaches.

WAAS (GPS plus a correction signal transmitted by separate geostationary satellites) improves the resolution to CAT I standards - an "LPV" approach. However, it's still not sufficient for CAT II/III approaches.

You basically use the regular altimeter for altitude during such approaches. It's accurate enough to get to the decision height of a few hundred feet AGL after which you have to fly the rest of the approach manually anyway. That is true with CAT I ILS as well.

A CAT III ILS approach requires very specialized equipment and training because you basically fly it all the way down to the runway with no requirement for visibility until moments before touchdown.

However, GPS is a lot better when it is not being obstructed by buildings around you. Planes have the benefit of unobstructed paths to GPS satellites and would get a better quality and accuracy then you might get with a phone in a city. Also, using good antennas helps. So, the altitude accuracy would be fine mostly.

Fine mostly, until it’s not. Then maybe those last couple meters might count.

Fine as in common practice and the way aviation has been operating safely for the past few decades.

Airports served only by an RNAV (GPS) approach with LPV minimums have different alternate minimums (for planning purposes). That’s the only real distinction. LPV approaches are classified as non-precision due to an ICAO technicality over the lack of a ground transmitter at the airport. GLS approaches compensate for this.

To count as a precision approach (which ILS is), the equipment and procedure has to meet certain accuracy requirements (e.g. <x> feet error at <y> distance/height from touchdown) as well as the ability for the equipment itself to alert the user that things are out of tolerance within <z> seconds of being detected faulty.

I.e. the transmitter on the ground must be able to detect that something is wrong and immediately cut out the signal / report it as inop to the receiver on the plane. When you're seconds from hitting the ground, that is a requirement.

Although GPS based approaches have some aspects of this, I believe that they just don't meet the exact requirements as written currently, although they are close.

Outside of the U.S., there are GPS-based precision approaches. I say GPS-based because they all require SBAS (space-based augmentation systems, like WAAS or EGNOS) rather than being pure GPS systems.

ICAO recognizes SBAS Cat I as a precision approach, and there are dozens of them in Europe (France, Austria at least).

IIRC GPS LPV is at best considered comparable to ILS Cat1, and GPS altitude data is pretty much forbidden to be used for that (radio altimeter is required for any kind of low-visibility approach, iirc, plus it's required for TAWS).

In good conditions the approaches might be indistinguishable, especially given that GPS approach once on final will be usually a copy of GPS one, but the safety of GPS LPV is considered much lower, for a good reason.

Can't edit anymore, but of course I meant that GPS LPV approach on final will be a copy of ILS :)

It is not uncommon to get +/- 15m error on the vertical axis. There is a significant delay in the signal too, so can't be used for landing.

This is why a recently calibrated altimeter is used for the altitude axis on GPS approaches. They also must have a constrained error margin to be certified, and it is factored into the buffered region around the approach.

I think you're being downvoted because the imprecision you speak of is a result of: consumer GPS receivers having a limited polling rate, limited precision, and only using one GPS frequency. You can't extrapolate the limitations of GPS on say a phone, to limitations on a commercial plane.

See the following from gps.gov: it can be accurate to within centimeters and with a high polling rate, but most consumer devices don't employ the necessary tricks. https://www.gps.gov/systems/gps/performance/accuracy/

Military gets more time resolution as well as no deliberate clock skew. Airliners won't have military GPS. Gps receivers for surveying require on the order of hours to resolve to "centimeters" using civilian GPS.

The highest resolution GPS receiver I have still has +/-15ns jitter, which over the course of a day is good enough for maybe 12-20cm resolution, but the first hour the track is wild, 50-80 meters off in varying directions.

My GPS is used as a time stamp source for lightning detection and location, so in aggregate with other detectors it can pinpoint a strike to within a meter or so anywhere in my hemisphere.

Reply on news.ycombinator.com