[clug] GPS, GIS, GTS, SDI and FOSS

[Scott Ferguson]

On 03/03/11 14:49, Brad Hards wrote:
> On Thursday, March 03, 2011 10:34:32 am Scott Ferguson wrote:
>> On 03/03/11 07:21, Brad Hards wrote:
>>> On Thursday, March 03, 2011 12:27:51 am Scott Ferguson wrote:
>>>> * Another solution to ionospheric distortion is internet sourced
>>>> corrective data - any suggestions/experience there?
>>>> A local base station is unlikely to work due to the local topography.
>>>
>>> I wouldn't have expected that there would be much variation in
>>> ionospheric delay across the ACT (or anywhere roughly north / south).
>>> The variation is strongest at dawn / dusk. Can you just put something on
>>> the roof?
>>>
>>> Brad
>>
>> Even that tiny variation adds significant error. The best non-military
>> GPS systems on the market cannot achieve less than 10m absolute accuracy
>> without corrective data. I can't quickly locate the figures - but after
>> local signal reflections the largest error factor is ionospheric
>> distortion (40%?). Of course the DSE sales rep will disagree with me... :-)
> The local signal reflections (multipath) are difficult to assess without seeing 
> the environment, so in best academic tradition, I'll assume those are zero.  
> Install choke rings around the antenna. Stay away from aircraft hangars and 
> Stobie poles. 
> 

I mention a primary site for convenience, and because it's the first
site we'll deploy *this* system - I am already working with a commercial
system in other locations, but it has proven, um, less than optimal.
The primary site is very hilly, has abysmal 3G reception (gotta love the
ACT) but does have lots of open sky. The open sky will decrease in
coming years - but that's a problem I'll deal with later, if my client's
still have the contract then.
While I'll consider assembling, or paying for someone to assemble a GPS
reciever - I'm intending to go with off-the shelf antenna.
Snow/DZ is not the client, so I'll be staying away from hangars... power
poles is not something I can avoid on other sites. :-/

> The ionosphere does all kinds of bad things to GPS signals, but the primary 
> effect is delay change, which adds error to the pseudo-range measurements. 
> 
> [For those playing along at home, GPS is essentially a trilateration system - 
> four distance measurements from satellites (in known positions) to a receiver 
> gives you a 3-D position plus an accurate time. Those measurements are 
> basically transmission time, which you turn into a distance by multiplying by 
> an estimate of the velocity of transmission (roughly the speed of light). The 
> ionosphere makes it hard to guess what that velocity is though.]
> 
>> I was probably (late night post) a little unclear - my requirements are
>> for reproducible 1/2 meter to 1 metre absolute accuracy in the ACT with
>> sub-2 minute warm starts at less than $1K per unit.
> In real time? Are you OK with post-processing in the corrections?

Part of the requirements is to be able to log data for thousands of
plants, many of which are less than 3 metres apart - and much of the
data collection will be done by semi-skilled labourers with minimal
training so I need to make the system as accurate and intuitive as
possible - basically a satellite image with waypoints imported from the
architectural plans, and populated by the botanical database. We need to
be certain that the operators aren't confused when sent to specific
locations, and that moisture levels and other data is not wrongly
attributed.

Depends how slow the post processing is... and how regularly the
correction data has to be updated. My intention is to use netbook
coupled to the GPS units. Currently trialling an ASUS PC Eee 701SD so
mild crunching shouldn't be a problem - and any delays are better that
the paper based reporting system required by the client's client.

The primary need for the system is one particular maintenance contract
the client holds - but we would like to make the system flexible enough
to use on other jobs, in other locations. While there are many GPS
location/mapping/data logging systems on the market (and in use by
Telstra, ACTEWAGL etc) - they are too expensive for smaller clients, and
too closed to be truly useful.

I suspect much of what we want to do for the primary site would interest
people like Jim Croft and others. I also believe it would add value to
my client's contract - and while my client is not IT savvy they kind of
grasp my explanation of FOSS and the benefits. It helps that they've
also sample some of the limitations of the closed offerings... eg.
solutions that will take 5 years to break even - presuming they just
work, require no maintenance, and the hardware and software concerned
doesn't reduce in cost during that period giving competitors an advantage.

> 
>> http://en.wikipedia.org/wiki/Error_analysis_for_the_Global_Positioning_Syst
>> em#Atmospheric_effects
>>
>> Not sure what you mean by "put something on the roof"... antennae?  If
>> so it is a means of reducing signal reflection - doesn't correct for
>> lens effects.
> If you calculate the pseudo-ranges from a known location (say, the roof of 
> your office), and can figure out the error in each pseudo-range, then you can 
> just post-process that error out later, since most of the ACT will see 
> basically the same ionospheric effects, especially if you can work during the 
> nice parts of the day (or night).

Plenty of survey points on the primary site, for other locations in the
ACT I suspect Google and other maps might be sufficient given 1m
resolution is acceptable.
System needs to be operational 24/7, though will mostly be in use during
daylight hours. Aside from the primary site where the data logging is
mostly at fixed waypoints during the daytime, on other sites it could be
used for logging feral animals - which is done at night (same client,
different contract)

> 
> It turns out that there are some services / data freely available:
> http://igscb.jpl.nasa.gov/components/prods.html
> http://www.ga.gov.au/geodesy/sgc/wwwgps/

Excellent - very, very useful.

> 
> So if you can post-process guess the problem is back to a receiver that will 
> log the pseudo-ranges, and some software to apply the corrections.
> 
> Brad
> 

Agreed, which is why I see the receiver selection as my starting point.

While I've been subscribing to this list for a dozen or so years I
usually don't read it on a daily basis. Like a few lists I subscribe to
it's proven an invaluable source of searchable information. Occasionally
I'll ask a list a question, and then monitor the list for a few days
afterwards so I can make a few posts and try to return the favour. In
this particular instance I'm *especially* pleased I did.

Thank you (all)