locating via GSM, revisited

[joerg at openmoko.org]

Am So  20. April 2008 schrieb andrzej zaborowski:
> On 20/04/2008, joerg at openmoko.org <joerg at openmoko.org> wrote:
> >  There is not so wellknown further more detailed information you may get 
from
> >  your GSM-modem (MobileStation, MS), which consists of:
> >  a) The (usually) 6 next nearby BTS (to be correct: next best RF-signal 
BTS),
> >  b) The distance to your active BTS, in increments of 550m (Timing 
Advance, TA)
> >  This additional info may be used to dramatically improve the precision of
> >  GSM-based location data.
> 
> Wow, didn't know that.  I think the timing advance must be a
> understood as the virtual distance the signal has travel, i.e. signal
> strength and not the distance we're interested in.  Or might a GSM
> modem have a way to know it's physical distance from the BTS? (Maybe
> the BTS can know, comparing the Signal-quality the phone sees and the
> quality of the signal *from* the phone to BTS, or even talking to
> other BTSes)


TA is generated (decided upon) by the BTS. The purpose is to have all the MS 
transmissions arrive right in time for their slot on the BTS, for there all 
MS share the same frequency on a timeslot scheme. So if a MS is more far away 
from BTS, it has to send more early to compensate for time the signal needs 
to travel to the BTS. The BTS can tell whether the signal is early or late 
and will advice the MS to compensate. That's TA value. The increment is a 1 
bitframe timeunit, what makes for 550m at speed of light.


> 
> It would be interesting to see how the three values: RF SQ, 

RF SQ is some more abstract value about the amount of redundancy that shall be 
imprinted on the datastream to help avoid dataloss, I guess. It's NOT closely 
related to actual physical distance, but tells a lot about interference etc.


> TA and GPS 
> distance correlate.

TA and GPS distance correlate 1:1, as long as the transmission from MS to BTS 
(and vice versa) is a straight line. TA just is the time the RF takes to 
travel to and fro BTS.



> >  From this data, we see it's quite possible to determine location to a
> >  precision of around 100 x 100m or even better.
> >  Of course this depends on the density of BTS again.
> >
> >
> >  To use this approach with GTA02 or GTA04/Diversity, it has to be 
evaluated
> >  whether we can get he NCEL-list from our GSM-modems.
> 
> Yes, we can, the info at
> 
http://wiki.openmoko.org/wiki/GTA01_gsm_modem#Neighbour_Cell_Information_.282.2C3.29
> came from my GTA01 modem but it works also on the GTA02 modem.  In
> 
http://wiki.openmoko.org/wiki/GTA01_gsm_modem#Serving_Cell_Information_.282.2C1.29
> we also get the Timing Advance value (called tav).  We need some stats
> :)

Fine! Pleased to hear this :-)


> 
> I'm not sure about forcing a reselection of the station but my guess
> would be that there is a command for that also.


If there's no such command already for it, we might probably be able to hack 
it to the baseband firmware (OM may do this I guess). 
Andy Green mentioned he thinks he might have seen such a command already. I 
will have some long nights to study the specs for GTA02 GSM modem, so I can 
tell for sure. Wouldn't mind to get a pointer instead.. ;-)

cheers
jOERG