[time-nuts] GPS position survey

[Poul-Henning Kamp]

In message <518733C4.1020705 at gmail.com>, Sarah White writes:
>On 5/6/2013 12:29 AM, Mark Sims wrote:

>> The 48-hour precision survey in Lady Heather uses a statistical
>>weighted median filter to arrive at its final location instead of
>>a simple average of fixes.   It processes data of one minute, hour,
>>and overlapping 24 hour intervals to calculate the final position.
>>It can produce a location that is 3-10 times better than simple
>>averaging of the fixes.

You can probably do even better than that:  Weigh the summing by DOP.

That gave me a factor two or so, because it reduced the weight of
samples suffering from multipath and other ailments.

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.