Dennis Lee Bieber

>On Wed, 12 Dec 2018 22:36:49 -0000, "NY"  declaimed the
>following:
>
>>
>>That's interesting. I'd have thought that with a good aerial and no
>>obstructions so you can see lots of satellites you should be able to get
>>very good results.
>>
>        For ground speed, not air speed... Having a 200kt ground speed with a
>100kt tail wind means an air speed of only 100kt.

Yes true, but pilots are well aware of wind speed and direction, meteo.
Simple addition is not too much for them..

So A GPS SOG (speed over ground) can be used to verify pitot tube type sensors.

I have done some research or better say experiments myself making ultrasonic
air speed  sensors,
basically for boating, some I posted about in sci.electronics.design.
These days you can buy those for maybe a hundred Euro or so.

All those methods combined is the LEAST that should be mounted in a many
millions dollar worth of plane.



>>My phone varies between about 20m and 5m radius of error for lateral
>>movement. Not sure how accurate it is for altitude. I sometimes find that I



>        I believe convention is to assume vertical error is 150% of lateral.
>Your phone is also most likely using just the C/A signal; airlines may be
>able to get decryption keys allowing use of the precision signal. I don't
>even know if any consumer level GPS is able to make use of dual frequency
>signals yet -- the precision signal is sent over two (or now, maybe four)
>frequencies, allowing the receiver to correct for ionospheric delays. C/A
>traditionally was sent on one frequency only.


Yes, for absolute position.
These days however units with GPS, GLONASS, Beidu are just 25 $.
Put 3 in a box and you can perhaps get smaller errors, Galileo is coming too,
if not already here.

All is relative, here I fly my Hubsan drone with 100% GPS controlled auto-pilot
to a few centimeter accuracy:
 https://www.youtube.com/watch?v=bI_0mjwlvNw

asm code is here:
 http://panteltje.com/panteltje/quadcopter/index.html

I mean even for a pressure based altimeter you need to calibrate at ground.
Maybe it is not so hard to (for example via radio) calibrate the on board GPS
relative to one on the landing strip,
basically what I do here.
Over short time scales the GPS position does not normally wander that much.
The math is very simple, see the C source code for composing the flight path on
same page:
 fly_waypoints-0.4.tgz

--- SoupGate-Win32 v1.05