Re: GPS sensitivity

[Stefan Monnier]

> agps-onlinec -c full -u [EMAIL PROTECTED] -k Fvhcv -la 35.22 -lo -79.46 -al
> 250 -p 100.00 &
[...]
> provided the location data prefed to agps-onlinec and the current time
> (system clock) are accurate enough. (clock doesn't need to be
> second-precise, but if it's off several hours or more then you've got

If you have some kind of regular IP connectivity, then your clock offset
should easily stay well within 1s (using an ntp daemon).


Stefan


PS: BTW, thank you all for your answers.  I now see that my experience
with my FR is simply normal so I shouldn't try and waste any time
figuring out what might be wrong with it.  I don't even need aGPS for my
intended use of the GPS (mostly for OSM).


___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: GPS sensitivity

[rakshat hooja]

In open sky conditions in Jaipur, India (outdoors) the Freerunner normally
gives me a 40s (+ - 10s) tiff while my N 95 takes 60s (+ - 10) [ cold starts
after rebooting on both ]. I have not timed them indoors but will do so and
report after Diwali

Rakshat


-- 
--
Please use Firefox as your web browser. Its protects you from spyware and is
also a very feature rich browser.
www.firefox.com
___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: GPS sensitivity

[Carl Lobo]

GPS works very well here. I've used it in Maharashtra and Goa.

On Tue, Oct 28, 2008 at 10:19 AM, Nishit Dave <[EMAIL PROTECTED]> wrote:
> On Mon, Oct 27, 2008 at 10:00 PM, Gora Mohanty <[EMAIL PROTECTED]> wrote:
>>
>> On Mon, 27 Oct 2008 21:30:22 +0530
>> "Nishit Dave" <[EMAIL PROTECTED]> wrote:
>> [...]
>> > I recently tested the performance of a Nokia E71 with the FR, while
>> > standing
>> > 2 feet from a window inside my office building.
>> [...]
>>
>> Um, does the Nokia E71 have assisted GPS (AGPS), as I
>> rather suspect it does? If so, you are comparing apples
>> and oranges.
>>
>> Regards,
>
> Yes, it does, but do you think AGPS reliably works in India?  Google
> Mylocation service has a radius of 5km here.  The E71 had pinpointed our
> position very accurately, and I doubt if AGPS would have that level of
> triangulation accuracy.
>
> ___
> Openmoko community mailing list
> community@lists.openmoko.org
> http://lists.openmoko.org/mailman/listinfo/community
>
>

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: GPS sensitivity

[Nishit Dave]

On Mon, Oct 27, 2008 at 10:00 PM, Gora Mohanty <[EMAIL PROTECTED]> wrote:

> On Mon, 27 Oct 2008 21:30:22 +0530
> "Nishit Dave" <[EMAIL PROTECTED]> wrote:
> [...]
> > I recently tested the performance of a Nokia E71 with the FR, while
> standing
> > 2 feet from a window inside my office building.
> [...]
>
> Um, does the Nokia E71 have assisted GPS (AGPS), as I
> rather suspect it does? If so, you are comparing apples
> and oranges.
>
> Regards,
>

Yes, it does, but do you think AGPS reliably works in India?  Google
Mylocation service has a radius of 5km here.  The E71 had pinpointed our
position very accurately, and I doubt if AGPS would have that level of
triangulation accuracy.
___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: GPS sensitivity

[Joel Newkirk]

On Tue, 28 Oct 2008 01:36:36 +0530, Gora Mohanty <[EMAIL PROTECTED]> wrote:
> On Mon, 27 Oct 2008 20:56:54 +0100
> David Garabana Barro <[EMAIL PROTECTED]> wrote:
> [...]
>> I allways though Freerunner had AGPS. That's what wiki says:
>>
>> http://wiki.openmoko.org/wiki/Neo_FreeRunner_Hardware#AGPS
>>
>> Isn't it true?
> [...]
> 
> Yes, I believe that it does have AGPS capabilities. But AGPS
> needs someone sending out the AGPS data. I don't know that
> that is being done with OpenMoko.
> 
> Regards,
> Gora

http://wiki.openmoko.org/wiki/AGPS

http://svn.openmoko.org/developers/matt_hsu/agps-online/

aGPS works nicely but isn't fully utilized ATM in most cases. You can
register for a free account at u-blox.com (chipset manufacturer for the GPS
in the FR - send empty email to agps-account at u-blox.com and get password
back) and use a tool called 'agps-onlinec' (at the SVN above) in the
instance I tested, and feed it some relevant data:

agps-onlinec -c full -u [EMAIL PROTECTED] -k Fvhcv -la 35.22 -lo -79.46 -al
250 -p 100.00 &

the -u is email address used to register above, '-k' is the password
provided in the reply email, '-la' and '-lo' and '-al' are current lat and
long and altitude, and the '-p' is precision in kilometers. If you have a
very accurate position already (eg you're at home and know precisely what
the GPS /should/ read) you can get a fix in a very short time with "-p 1",
provided the location data prefed to agps-onlinec and the current time
(system clock) are accurate enough. (clock doesn't need to be
second-precise, but if it's off several hours or more then you've got
problems)  On warm restart of the GPS it retains this data within the GPS,
on cold restart it is lost and must be resupplied from outside or it takes
significantly longer again to get a fix.  

Frameworkd (FSO) apparently stores almanac data automatically and feeds it
back into the GPS on startup, so if you're within a couple hours of
previous GPS usage (I presume) you get a faster fix.  Ideal in my mind
would be to check cell tower ID and/or wifi AP MAC against
previously-cached records and feed the location data from those records
(with appropriate precision data, perhaps 10km for cell-tower?) into
agps-onlinec and let it retrieve data via GPRS from the assistance server. 
(even better would be a local service ON the FR that retrieves the full
2-week dataset and serves up the appropriate data on demand, but I'm not
even sure if that can be done without careful dissection of both protocols
- the long-term data isn't supported by the GPS inside the FR)

j



___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: GPS sensitivity

[nick d.]

Online mode: http://wiki.openmoko.org/wiki/Neo_FreeRunner_GPS#Assist_Online

I've signed up for an account but not sure if the scripts worked for me or
not. Haven't tested it fully.

Make sure you update the wiki if you have any success or more information.

-Nick

On Tue, Oct 28, 2008 at 6:24 AM, Johny Tenfinger <[EMAIL PROTECTED]>wrote:

> > So, given that both the Neo 1973, and the Freerunner apparently
> > have AGPS capabilities, I think that what needs to be figured
> > out is how to send out AGPS data from community servers. Anyone
> > with any thoughts on this?
>
> Neo1973 doesn't have AGPS capability, but FreeRunner has, and some
> time ago on this mailist was program to do GPS assistance in "online
> mode" - I don't know if it works, but when I have fix and I'm doing
> warm reset of FR GPS, it gets fix in ~5 seconds, so I'm think this
> result is also possible with working AGPS.
> dos
>
> ___
> Openmoko community mailing list
> community@lists.openmoko.org
> http://lists.openmoko.org/mailman/listinfo/community
>
___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: GPS sensitivity

[Johny Tenfinger]

> So, given that both the Neo 1973, and the Freerunner apparently
> have AGPS capabilities, I think that what needs to be figured
> out is how to send out AGPS data from community servers. Anyone
> with any thoughts on this?

Neo1973 doesn't have AGPS capability, but FreeRunner has, and some
time ago on this mailist was program to do GPS assistance in "online
mode" - I don't know if it works, but when I have fix and I'm doing
warm reset of FR GPS, it gets fix in ~5 seconds, so I'm think this
result is also possible with working AGPS.
dos

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: GPS sensitivity

[Gora Mohanty]

On Mon, 27 Oct 2008 21:04:55 +0100
"Leonti Bielski" <[EMAIL PROTECTED]> wrote:

> We are talking here about hardware GPS performance of Freerunner.
> So to compare two GPS devices they have to be in equal positions. 
[...]

Yes, fair enough. You are right. Please see my earlier reply
to "Johny Tenfinger". However, for an end-user, the question
becomes whether the device quote just works unquote.

So, given that both the Neo 1973, and the Freerunner apparently
have AGPS capabilities, I think that what needs to be figured
out is how to send out AGPS data from community servers. Anyone
with any thoughts on this?

Regards,
Gora

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: GPS sensitivity

[Gora Mohanty]

On Mon, 27 Oct 2008 20:50:01 +0100
"Johny Tenfinger" <[EMAIL PROTECTED]> wrote:

> > Again, I must ask. Is AGPS enabled on the Nokia N95?
> 
> But when you want to compare N95 GPS with FreeRunner GPS, you must
> turn AGPS off in Nokia, or turn AGPS on in FreeRunner...
[...]

You are right. Sorry, I conflated two things in my various
replies:
(a) One is an apples-to-apples comparison: Turn off AGPS, and
look at the capabilities of the internal GPS receivers. I
hope to soon do a systematic study of the errors of various
devices by sitting in the sun for a couple of hours.
(b) Compare device to device (apples to oranges), by turning on
the best capabilities of each. Here I believe that an AGPS
device must perform as well, or typically better.

Regards,
Gora

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: GPS sensitivity

[Gora Mohanty]

On Mon, 27 Oct 2008 20:56:54 +0100
David Garabana Barro <[EMAIL PROTECTED]> wrote:
[...]
> I allways though Freerunner had AGPS. That's what wiki says:
> 
> http://wiki.openmoko.org/wiki/Neo_FreeRunner_Hardware#AGPS
> 
> Isn't it true?
[...]

Yes, I believe that it does have AGPS capabilities. But AGPS
needs someone sending out the AGPS data. I don't know that
that is being done with OpenMoko.

Regards,
Gora

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: GPS sensitivity

[Leonti Bielski]

We are talking here about hardware GPS performance of Freerunner.
So to compare two GPS devices they have to be in equal positions. If Neo
Freerunner not using aGPS N95 shouldn't to for results to be comparable. I
didn't think I needed to clarify this.
So to answer you question - no, Nokia were not using aGPS, with it it gets
fixed in 12-15 seconds. But without it it's no better than Freerunner. I did
not say it's worse, but it's not better either.
Leonti

On Mon, Oct 27, 2008 at 8:43 PM, Gora Mohanty <[EMAIL PROTECTED]> wrote:

> On Mon, 27 Oct 2008 20:09:34 +0100
> "Leonti Bielski" <[EMAIL PROTECTED]> wrote:
>
> > My brother has Nokia N95 and it has about the same performance as my Neo.
> > No fix inside, around 40 seconds fix outside.
>
> Again, I must ask. Is AGPS enabled on the Nokia N95? That needs
> to be specifically turned on, and a network connection available.
> I only had the chance to play with a Nokia N95 for a few hours,
> but with AGPS on, I got a sync indoors (well, not too far from a
> window, but in an urban canyon with very poor line of sight) within
> 30s. Most GPS receivers I have tried there have not found a sync.
>
> > So it's not worse than others.
> [...]
>
> Much as we all love OpenMoko, I find it hard to believe that a
> device with AGPS has comparable or worse performance than one
> without it. There are certainly many things about the N95 that
> suck, but I sincerely doubt that it has worse GPS location
> capabilities.
>
> Regards,
> Gora
>
> ___
> Openmoko community mailing list
> community@lists.openmoko.org
> http://lists.openmoko.org/mailman/listinfo/community
>
___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: GPS sensitivity

[David Garabana Barro]

On Monday 27 October 2008 20:43:43 Gora Mohanty wrote:

>
> Much as we all love OpenMoko, I find it hard to believe that a
> device with AGPS has comparable or worse performance than one
> without it. There are certainly many things about the N95 that
> suck, but I sincerely doubt that it has worse GPS location
> capabilities.

I allways though Freerunner had AGPS. That's what wiki says:

http://wiki.openmoko.org/wiki/Neo_FreeRunner_Hardware#AGPS

Isn't it true?

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: GPS sensitivity

[Johny Tenfinger]

> Again, I must ask. Is AGPS enabled on the Nokia N95?

But when you want to compare N95 GPS with FreeRunner GPS, you must
turn AGPS off in Nokia, or turn AGPS on in FreeRunner...

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: GPS sensitivity

[Gora Mohanty]

On Mon, 27 Oct 2008 20:09:34 +0100
"Leonti Bielski" <[EMAIL PROTECTED]> wrote:

> My brother has Nokia N95 and it has about the same performance as my Neo.
> No fix inside, around 40 seconds fix outside.

Again, I must ask. Is AGPS enabled on the Nokia N95? That needs
to be specifically turned on, and a network connection available.
I only had the chance to play with a Nokia N95 for a few hours,
but with AGPS on, I got a sync indoors (well, not too far from a
window, but in an urban canyon with very poor line of sight) within
30s. Most GPS receivers I have tried there have not found a sync.

> So it's not worse than others.
[...]

Much as we all love OpenMoko, I find it hard to believe that a
device with AGPS has comparable or worse performance than one
without it. There are certainly many things about the N95 that
suck, but I sincerely doubt that it has worse GPS location
capabilities.

Regards,
Gora

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: GPS sensitivity

[Stefan Strahl]

Good evening Freerunners :)

I can report a good GPS performance :) Two weeks ago I was testing
tangogps on a hike. A friend was also tracking using his new Garmin.
See http://www.student.uni-oldenburg.de/stefan.strahl/FreerunnerVSGarmin.jpg
for a screenshot. Red is the Garmin track, blue the Freerunner
recordings. The facts are:

- the Freerunner needed several minutes until he got a fix and the
first positions were wrong (wrong start is labeled "-0", we started at
the railway station / lower left corner of the screenshot)
- the Garmin showed after some time a drift (reflections from the
mountains?) of which interestingly the freerunner did not suffer from
(we definitely walked along the blue track).
- the battery lifetime of the Garmin is much longer ;) - my batteries
were empty after ~4 hours

So the GPS accuracy is quite good, nothing to complain about. And once
someone found time to implement the A-GPS functionality we shouldn't
worry about the first-fix times. So this problem is resolvable in
finite time... ;)

Happy freephoning and -running

:) stefan

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: GPS sensitivity

[Leonti Bielski]

My brother has Nokia N95 and it has about the same performance as my Neo.
No fix inside, around 40 seconds fix outside.
So it's not worse than others.
Also I have good bluetooth gps, and it has problems getting fix in the city
area or while moving, so you can't blame Neo for not getting fix - it's
other conditions (providing you have software or hardware workarounds).
Leonti
___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: GPS sensitivity

[Gora Mohanty]

On Mon, 27 Oct 2008 21:30:22 +0530
"Nishit Dave" <[EMAIL PROTECTED]> wrote:
[...]
> I recently tested the performance of a Nokia E71 with the FR, while standing
> 2 feet from a window inside my office building. 
[...]

Um, does the Nokia E71 have assisted GPS (AGPS), as I
rather suspect it does? If so, you are comparing apples
and oranges.

Regards,
Gora

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: GPS sensitivity

[Stefan Fröbe]

> The FR still takes some minutes to get a fix when in a moving vehicle, and
> anywhere inside a building, it is as if it never worked.
>
> Nothing new
>

Speaking from daily experience the FR is in no way less sensitive than any
other GPSr I have used or was around with while geocaching, and that is my
main reason I bought it! Actually just this weekend I was caching in a city,
and while reception was generally very good, when I noticed a drift in one
spot the Garmin Colorado my colleague had with her showed the same erratic
behaviour.

And with recent SW, first-fix-times have gone down immensely: I hardly ever
need more than 40s, and when standing still I can barely set up tangogps
before I have a signal - moving on a bike is a different story, but putting
it upright in the pocket is the key here.
This is all with 2008.9 and newer kernel, HW fix applied manually.

Stefan
___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: GPS sensitivity

[Iain B. Findleton]

I find the FR more or less equivalent to other GPS devices I have. The
biggest factor I notice is weather conditions. In overcast conditions I
can use the GPS anywhere inside my house, as well as outside. Under
clear sky conditions, the GPS will only work outside and will not
acquire a first fix unless I am more or less in the open, away from
trees, houses, etc.

Some people claim that there are newer GPS chips that work better.
Perhaps its the nature of the beast.

Nishit Dave wrote:
> On Mon, Oct 27, 2008 at 9:18 PM, Alastair Johnson
> <[EMAIL PROTECTED] > wrote:
>
> Stefan Monnier wrote:
> > The FR is the first device I use with a GPS, so I don't know what's
> > considered as "normal" w.r.t GPS function.  I find that my FR's GPS
> > never works inside a building (e.g. at home), and even outside
> in the
> > streets of Montreal, it seems to only be able to get a first fix
> if I'm
> > in a somewhat open area (i.e. not in a street but on a place, in
> > a park), and also it seems to rarely if ever be able to get a
> first fix
> > when it's in my pocket.
> >
> > Is that normal?  My FR does have the capacitor in the µSD slot
> and it
> > has a fairly recent kernel (don't know if that means it has the
> software
> > fix that "stops the µSD clock when possible", does it?)
>
> It's a little tricky to describe 'normal' since the movement of
> the sats
> gives some inherent variability. Getting the first fix also requires
> significantly more signal than maintaining a fix once acquired, and it
> seems to help being stationary when doing it. I don't know how limited
> the view of the sky is in Montreal, but 'urban canyon' effects
> have long
> been a problem for GPS systems due to limited view of the sky
> (can't see
> enough sats) and multiple reflected signals. That said, since the SD
> clocking fixes were added to the kernel I find the Freerunner
> usually at
> worst as quick as my Garmin Gecko at getting a fix, and substantially
> better at keeping it. The Freerunner will often keep a fix indoors
> when
> the Gecko hasn't a hope. OTOH the Gecko is hardly state of the art
> now,
> so expectations may be different.
>
>  
> I recently tested the performance of a Nokia E71 with the FR, while
> standing 2 feet from a window inside my office building.  Most of the
> view (line of sight) was obstructed by another large building 200 ft
> away, although the sky over it can be viewed from the said window.
>
> The Nokia got a fix and started to download maps with google maps in
> 15 seconds.  I was unsuccessful in getting the FR to register a fix at
> the same position for greater than 15 minutes, after which the
> experiment was stopped. Qtextended 4.4.1, MappingDemo used.  The FR
> version I use is from the time when the software fix to the hardware
> problem was about to be released.  The FR still takes some minutes to
> get a fix when in a moving vehicle, and anywhere inside a building, it
> is as if it never worked. 
>
> Nothing new.
> 
>
> ___
> Openmoko community mailing list
> community@lists.openmoko.org
> http://lists.openmoko.org/mailman/listinfo/community
>   


-- 
Iain B. Findleton
Tel: 514-457-0744



___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: GPS sensitivity

[Nishit Dave]

On Mon, Oct 27, 2008 at 9:18 PM, Alastair Johnson <[EMAIL PROTECTED]>wrote:

> Stefan Monnier wrote:
> > The FR is the first device I use with a GPS, so I don't know what's
> > considered as "normal" w.r.t GPS function.  I find that my FR's GPS
> > never works inside a building (e.g. at home), and even outside in the
> > streets of Montreal, it seems to only be able to get a first fix if I'm
> > in a somewhat open area (i.e. not in a street but on a place, in
> > a park), and also it seems to rarely if ever be able to get a first fix
> > when it's in my pocket.
> >
> > Is that normal?  My FR does have the capacitor in the µSD slot and it
> > has a fairly recent kernel (don't know if that means it has the software
> > fix that "stops the µSD clock when possible", does it?)
>
> It's a little tricky to describe 'normal' since the movement of the sats
> gives some inherent variability. Getting the first fix also requires
> significantly more signal than maintaining a fix once acquired, and it
> seems to help being stationary when doing it. I don't know how limited
> the view of the sky is in Montreal, but 'urban canyon' effects have long
> been a problem for GPS systems due to limited view of the sky (can't see
> enough sats) and multiple reflected signals. That said, since the SD
> clocking fixes were added to the kernel I find the Freerunner usually at
> worst as quick as my Garmin Gecko at getting a fix, and substantially
> better at keeping it. The Freerunner will often keep a fix indoors when
> the Gecko hasn't a hope. OTOH the Gecko is hardly state of the art now,
> so expectations may be different.
>

I recently tested the performance of a Nokia E71 with the FR, while standing
2 feet from a window inside my office building.  Most of the view (line of
sight) was obstructed by another large building 200 ft away, although the
sky over it can be viewed from the said window.

The Nokia got a fix and started to download maps with google maps in 15
seconds.  I was unsuccessful in getting the FR to register a fix at the same
position for greater than 15 minutes, after which the experiment was
stopped. Qtextended 4.4.1, MappingDemo used.  The FR version I use is from
the time when the software fix to the hardware problem was about to be
released.  The FR still takes some minutes to get a fix when in a moving
vehicle, and anywhere inside a building, it is as if it never worked.

Nothing new.
___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community

Re: GPS sensitivity

[Alastair Johnson]

Stefan Monnier wrote:
> The FR is the first device I use with a GPS, so I don't know what's
> considered as "normal" w.r.t GPS function.  I find that my FR's GPS
> never works inside a building (e.g. at home), and even outside in the
> streets of Montreal, it seems to only be able to get a first fix if I'm
> in a somewhat open area (i.e. not in a street but on a place, in
> a park), and also it seems to rarely if ever be able to get a first fix
> when it's in my pocket.
> 
> Is that normal?  My FR does have the capacitor in the µSD slot and it
> has a fairly recent kernel (don't know if that means it has the software
> fix that "stops the µSD clock when possible", does it?)

It's a little tricky to describe 'normal' since the movement of the sats 
gives some inherent variability. Getting the first fix also requires 
significantly more signal than maintaining a fix once acquired, and it 
seems to help being stationary when doing it. I don't know how limited 
the view of the sky is in Montreal, but 'urban canyon' effects have long 
been a problem for GPS systems due to limited view of the sky (can't see 
enough sats) and multiple reflected signals. That said, since the SD 
clocking fixes were added to the kernel I find the Freerunner usually at 
worst as quick as my Garmin Gecko at getting a fix, and substantially 
better at keeping it. The Freerunner will often keep a fix indoors when 
the Gecko hasn't a hope. OTOH the Gecko is hardly state of the art now, 
so expectations may be different.

___
Openmoko community mailing list
community@lists.openmoko.org
http://lists.openmoko.org/mailman/listinfo/community