| Author |
Message |
TK
Guest
|
 Posted:
Fri Dec 09, 2005 5:08 pm Post subject:
Accuracy vs TTFF |
|
|
Hi all
I have started looking into this yesturday, so i may be talking
nonsense.
I am looking an embedded GPS module that would allow me to get a very
fast <10s TTFF.
Accuracy of location can be within 300-500m. is there any module that
will allow me to do this. is there some way i can force it not to look
for ephermeris data. Can some one can point me to similar project been
done before, or point me to right module.
cheers
TK |
|
| Back to top |
|
 |
Sam Wormley
Guest
|
| Posted:
Fri Dec 09, 2005 5:08 pm Post subject:
Re: Accuracy vs TTFF |
|
|
TK wrote:
| Quote: | Hi all
I have started looking into this yesturday, so i may be talking
nonsense.
I am looking an embedded GPS module that would allow me to get a very
fast <10s TTFF.
Accuracy of location can be within 300-500m. is there any module that
will allow me to do this. is there some way i can force it not to look
for ephermeris data. Can some one can point me to similar project been
done before, or point me to right module.
|
What good is a bogus (or very inaccurate PVT solution) is less than
ten seconds? One needs a minimum of 30 seconds to download current
ephemeris data, assuming it is not already current. Otherwise position
error could be a lot greater than 500m, if even meaningful. |
|
| Back to top |
|
|
TK
Guest
|
| Posted:
Fri Dec 09, 2005 5:08 pm Post subject:
Re: Accuracy vs TTFF |
|
|
yeah, you are right in a way. just thinking of ways to get the fastest
fix. need to get fix for animals that stick their head out of water for
only a few seconds. so looking for smallest , lightest device, but aslo
looking for a way to get a fix in aprox 10s or less. any ideas are most
welcome. |
|
| Back to top |
|
|
Marc Brett
Guest
|
| Posted:
Fri Dec 09, 2005 11:08 pm Post subject:
Re: Accuracy vs TTFF |
|
|
On 9 Dec 2005 08:00:24 -0800, "TK" <taimurkhan@gmail.com> wrote:
| Quote: | Hi all
I have started looking into this yesturday, so i may be talking
nonsense.
I am looking an embedded GPS module that would allow me to get a very
fast <10s TTFF.
Accuracy of location can be within 300-500m. is there any module that
will allow me to do this. is there some way i can force it not to look
for ephermeris data. Can some one can point me to similar project been
done before, or point me to right module.
|
You may want to look at Assisted GPS (A-GPS), which combines GPS technology with
assistance from the mobile phone network. TTFF can be less than a second.
Downside is that you need a mobile phone network with a backend server to get
this kind of performance.
Texas Instruments and Motorola (FS Oncore), and probably others, make A-GPS
modules. |
|
| Back to top |
|
|
jim_gps
Guest
|
| Posted:
Fri Dec 09, 2005 11:13 pm Post subject:
Re: Accuracy vs TTFF |
|
|
It is possible, and coincidently I've already done a similar project
(for different reasons).
The way we did it was to log just raw pseudo-ranges, and then at a
later date processed that data with the ephemeris data that was logged
at a reference receiver. The good news is that this resulted in a fix
being available in only a few seconds, the only downside is that the
actual position wasn't available in real time... you had to work it out
afterwards.
Thanks!
Jim
TK wrote:
| Quote: | yeah, you are right in a way. just thinking of ways to get the fastest
fix. need to get fix for animals that stick their head out of water for
only a few seconds. so looking for smallest , lightest device, but aslo
looking for a way to get a fix in aprox 10s or less. any ideas are most
welcome. |
|
|
| Back to top |
|
|
gpshelp
Guest
|
| Posted:
Fri Dec 09, 2005 11:34 pm Post subject:
Re: Accuracy vs TTFF |
|
|
| Quote: | Hi all
I have started looking into this yesturday, so i may be talking
nonsense.
I am looking an embedded GPS module that would allow me to get a very
fast <10s TTFF.
Accuracy of location can be within 300-500m. is there any module that
will allow me to do this. is there some way i can force it not to look
for ephermeris data. Can some one can point me to similar project been
done before, or point me to right module.
|
To get 300m-500m accuracy, the GPS module needs some way of calculating
the positions of the satellites to within about 100m. If you use the
current ephemeris data, the orbits will be accurate to within 5m, but
this quickly degrades outside the fit interval. This is why the GPS
system also provides almanac data, which is much less accurate but
degrades more slowly, and can be used for several weeks. However I
don't think the almanac would give you 100m accuracy. The key point is
that ephemeris data which is a week old would give orbits which were
out by much much more than a 100m, and I doubt that you get that
accuracy with ephemeris data which was more than six hours outside its
fit interval.
So if you want 10 second fixes you will need some way to get fresh
ephemeris to the GPS module. A lot of units are capable of accepting
ephemeris data from an external source. If you had a comms link you
could transmit it to the unit wirelessly. If not, the solutions get
much mor ecomplicated.
However, if you don't have a comms link, then you presumably are just
storing the data for later analysis. In that case, you don't need to
calculate the GPS positions in the field, you just need to acquire at
least 3-4 satellites, and log the pseudorange measurement for each
satellite. This should be possible in 10 seconds, but would depend
greatly on the GPS module. Once you have the pseduoranges, the position
could then be calculated after the data was downloaded, using ephemeris
data acquired from the internet.
Peter. |
|
| Back to top |
|
|
gpshelp
Guest
|
| Posted:
Fri Dec 09, 2005 11:34 pm Post subject:
Re: Accuracy vs TTFF |
|
|
| Quote: | Hi all
I have started looking into this yesturday, so i may be talking
nonsense.
I am looking an embedded GPS module that would allow me to get a very
fast <10s TTFF.
Accuracy of location can be within 300-500m. is there any module that
will allow me to do this. is there some way i can force it not to look
for ephermeris data. Can some one can point me to similar project been
done before, or point me to right module.
|
To get 300m-500m accuracy, the GPS module needs some way of calculating
the positions of the satellites to within about 100m. If you use the
current ephemeris data, the orbits will be accurate to within 5m, but
this quickly degrades outside the fit interval. This is why the GPS
system also provides almanac data, which is much less accurate but
degrades more slowly, and can be used for several weeks. However I
don't think the almanac would give you 100m accuracy. The key point is
that ephemeris data which is a week old would give orbits which were
out by much much more than a 100m, and I doubt that you get that
accuracy with ephemeris data which was more than six hours outside its
fit interval.
So if you want 10 second fixes you will need some way to get fresh
ephemeris to the GPS module. A lot of units are capable of accepting
ephemeris data from an external source. If you had a comms link you
could transmit it to the unit wirelessly. If not, the solutions get
much mor ecomplicated.
However, if you don't have a comms link, then you presumably are just
storing the data for later analysis. In that case, you don't need to
calculate the GPS positions in the field, you just need to acquire at
least 3-4 satellites, and log the pseudorange measurement for each
satellite. This should be possible in 10 seconds, but would depend
greatly on the GPS module. Once you have the pseduoranges, the position
could then be calculated after the data was downloaded, using ephemeris
data acquired from the internet.
Peter. |
|
| Back to top |
|
|
TK
Guest
|
| Posted:
Sat Dec 10, 2005 1:08 am Post subject:
Re: Accuracy vs TTFF |
|
|
thanks all for the info
i think the best option for me would be offline processing and just
storing the pseudo-locations, then use ofline proicessing useing
downloaded ephermeris. any idea which module would let me do it and how
?
@JIM do you have details of your project available. I am highly
interested in your implementation.
cheers all for your help |
|
| Back to top |
|
|
Guest
|
| Posted:
Sat Dec 10, 2005 1:08 am Post subject:
Re: Accuracy vs TTFF |
|
|
On Fri, 09 Dec 2005 16:10:52 GMT, Sam Wormley <swormley1@mchsi.com>
wrote:
| Quote: | TK wrote:
Hi all
I have started looking into this yesturday, so i may be talking
nonsense.
I am looking an embedded GPS module that would allow me to get a very
fast <10s TTFF.
Accuracy of location can be within 300-500m. is there any module that
will allow me to do this. is there some way i can force it not to look
for ephermeris data. Can some one can point me to similar project been
done before, or point me to right module.
What good is a bogus (or very inaccurate PVT solution) is less than
ten seconds? One needs a minimum of 30 seconds to download current
ephemeris data, assuming it is not already current. Otherwise position
error could be a lot greater than 500m, if even meaningful.
|
Some "Assisted GPS" (aGPS) systems send an ephemeris via land-based
comm links. There are as many implementations, a good primer is
http://www.gpsworld.com/gpsworld/article/articleDetail.jsp?id=12287
Just because Trimble and the US govt sites don't have the answer does
not mean it is not available. ;-). |
|
| Back to top |
|
|
Sam Wormley
Guest
|
| Posted:
Sat Dec 10, 2005 1:08 am Post subject:
Re: Accuracy vs TTFF |
|
|
kc wrote:
| Quote: | optional@telus.net writes:
On Fri, 09 Dec 2005 16:10:52 GMT, Sam Wormley <swormley1@mchsi.com
wrote:
TK wrote:
Hi all
I have started looking into this yesturday, so i may be talking
nonsense.
I am looking an embedded GPS module that would allow me to get a very
fast <10s TTFF.
Accuracy of location can be within 300-500m. is there any module that
will allow me to do this. is there some way i can force it not to look
for ephermeris data. Can some one can point me to similar project been
done before, or point me to right module.
What good is a bogus (or very inaccurate PVT solution) is less than
ten seconds? One needs a minimum of 30 seconds to download current
ephemeris data, assuming it is not already current. Otherwise position
error could be a lot greater than 500m, if even meaningful.
Some "Assisted GPS" (aGPS) systems send an ephemeris via land-based
comm links. There are as many implementations, a good primer is
http://www.gpsworld.com/gpsworld/article/articleDetail.jsp?id=12287
Just because Trimble and the US govt sites don't have the answer does
not mean it is not available. ;-).
I've seen assisted GPS on a cell phone get fixes consistently in about
7 seconds. Not only does the cellular netowrk feed the GPS ephemeris,
it also supplys a reference time (within 2 seconds helps a lot) and a
reference location (within about 30KM helps a lot). First time I saw
it I had to cycle the power on the unit to prove to myself it was
working from a cold start!
|
7 seconds from the time a placed call is connected?
7 seconds from when?
My phone is "on" most of the time in my pocket... probably allowing
the GPS chip to update itself periodically.
The ephemeris data is repeated every 30 seconds from the GPS satellites.
Are you saying that the ephemeris data is available from cell towers? |
|
| Back to top |
|
|
kc
Guest
|
| Posted:
Sat Dec 10, 2005 1:08 am Post subject:
Re: Accuracy vs TTFF |
|
|
optional@telus.net writes:
| Quote: | On Fri, 09 Dec 2005 16:10:52 GMT, Sam Wormley <swormley1@mchsi.com
wrote:
TK wrote:
Hi all
I have started looking into this yesturday, so i may be talking
nonsense.
I am looking an embedded GPS module that would allow me to get a very
fast <10s TTFF.
Accuracy of location can be within 300-500m. is there any module that
will allow me to do this. is there some way i can force it not to look
for ephermeris data. Can some one can point me to similar project been
done before, or point me to right module.
What good is a bogus (or very inaccurate PVT solution) is less than
ten seconds? One needs a minimum of 30 seconds to download current
ephemeris data, assuming it is not already current. Otherwise position
error could be a lot greater than 500m, if even meaningful.
Some "Assisted GPS" (aGPS) systems send an ephemeris via land-based
comm links. There are as many implementations, a good primer is
http://www.gpsworld.com/gpsworld/article/articleDetail.jsp?id=12287
Just because Trimble and the US govt sites don't have the answer does
not mean it is not available. ;-).
I've seen assisted GPS on a cell phone get fixes consistently in about |
7 seconds. Not only does the cellular netowrk feed the GPS ephemeris,
it also supplys a reference time (within 2 seconds helps a lot) and a
reference location (within about 30KM helps a lot). First time I saw
it I had to cycle the power on the unit to prove to myself it was
working from a cold start! |
|
| Back to top |
|
|
Sam Wormley
Guest
|
| Posted:
Sat Dec 10, 2005 1:08 am Post subject:
Re: Accuracy vs TTFF |
|
|
optional@telus.net wrote:
Thanks. |
|
| Back to top |
|
|
Dale DePriest
Guest
|
| Posted:
Sat Dec 10, 2005 9:08 am Post subject:
Re: Accuracy vs TTFF |
|
|
Sam Wormley wrote:
| Quote: | kc wrote:
optional@telus.net writes:
On Fri, 09 Dec 2005 16:10:52 GMT, Sam Wormley <swormley1@mchsi.com
wrote:
TK wrote:
Hi all
I have started looking into this yesturday, so i may be talking
nonsense.
I am looking an embedded GPS module that would allow me to get a very
fast <10s TTFF.
Accuracy of location can be within 300-500m. is there any module that
will allow me to do this. is there some way i can force it not to look
for ephermeris data. Can some one can point me to similar project been
done before, or point me to right module.
What good is a bogus (or very inaccurate PVT solution) is less than
ten seconds? One needs a minimum of 30 seconds to download current
ephemeris data, assuming it is not already current. Otherwise position
error could be a lot greater than 500m, if even meaningful.
Some "Assisted GPS" (aGPS) systems send an ephemeris via land-based
comm links. There are as many implementations, a good primer is
http://www.gpsworld.com/gpsworld/article/articleDetail.jsp?id=12287
Just because Trimble and the US govt sites don't have the answer does
not mean it is not available. ;-).
I've seen assisted GPS on a cell phone get fixes consistently in about
7 seconds. Not only does the cellular netowrk feed the GPS ephemeris,
it also supplys a reference time (within 2 seconds helps a lot) and a
reference location (within about 30KM helps a lot). First time I saw
it I had to cycle the power on the unit to prove to myself it was
working from a cold start!
7 seconds from the time a placed call is connected?
7 seconds from when?
My phone is "on" most of the time in my pocket... probably allowing
the GPS chip to update itself periodically.
The ephemeris data is repeated every 30 seconds from the GPS satellites.
Are you saying that the ephemeris data is available from cell towers?
|
Yes, that is one of the tricks. The ephemeris data can be transmitted
much more quickly from the cell phone than the 50 baud from the
satellite. I have heard of fixes in 5 seconds, certainly 7 is
reasonable. For example, you leave a building where all GPS signals are
blocked and with 2 or 3 seconds it has a fix since it already has the
setup data.
Dale
--
_ _ Dale DePriest
/`) _ // http://users.cwnet.com/dalede
o/_/ (_(_X_(` For GPS and GPS/PDAs |
|
| Back to top |
|
|
Sam Wormley
Guest
|
| Posted:
Sat Dec 10, 2005 5:08 pm Post subject:
Re: Accuracy vs TTFF |
|
|
Dale DePriest wrote:
| Quote: |
Sam Wormley wrote:
kc wrote:
optional@telus.net writes:
On Fri, 09 Dec 2005 16:10:52 GMT, Sam Wormley <swormley1@mchsi.com
wrote:
TK wrote:
Hi all
I have started looking into this yesturday, so i may be talking
nonsense.
I am looking an embedded GPS module that would allow me to get a very
fast <10s TTFF.
Accuracy of location can be within 300-500m. is there any module that
will allow me to do this. is there some way i can force it not to
look
for ephermeris data. Can some one can point me to similar project
been
done before, or point me to right module.
What good is a bogus (or very inaccurate PVT solution) is less than
ten seconds? One needs a minimum of 30 seconds to download current
ephemeris data, assuming it is not already current. Otherwise
position
error could be a lot greater than 500m, if even meaningful.
Some "Assisted GPS" (aGPS) systems send an ephemeris via land-based
comm links. There are as many implementations, a good primer is
http://www.gpsworld.com/gpsworld/article/articleDetail.jsp?id=12287
Just because Trimble and the US govt sites don't have the answer does
not mean it is not available. ;-).
I've seen assisted GPS on a cell phone get fixes consistently in about
7 seconds. Not only does the cellular netowrk feed the GPS ephemeris,
it also supplys a reference time (within 2 seconds helps a lot) and a
reference location (within about 30KM helps a lot). First time I saw
it I had to cycle the power on the unit to prove to myself it was
working from a cold start!
7 seconds from the time a placed call is connected?
7 seconds from when?
My phone is "on" most of the time in my pocket... probably allowing
the GPS chip to update itself periodically.
The ephemeris data is repeated every 30 seconds from the GPS
satellites.
Are you saying that the ephemeris data is available from cell towers?
Yes, that is one of the tricks. The ephemeris data can be transmitted
much more quickly from the cell phone than the 50 baud from the
satellite. I have heard of fixes in 5 seconds, certainly 7 is
reasonable. For example, you leave a building where all GPS signals are
blocked and with 2 or 3 seconds it has a fix since it already has the
setup data.
Dale
|
Thanks Dale!
-Sam |
|
| Back to top |
|
|
kc
Guest
|
| Posted:
Tue Dec 13, 2005 1:08 am Post subject:
Re: Accuracy vs TTFF |
|
|
Dale DePriest <Dale@gpsinformation.het> writes:
| Quote: | Sam Wormley wrote:
7 seconds from the time a placed call is connected?
7 seconds from when?
My phone is "on" most of the time in my pocket... probably allowing
the GPS chip to update itself periodically.
The ephemeris data is repeated every 30 seconds from the GPS satellites.
Are you saying that the ephemeris data is available from cell towers?
Yes, that is one of the tricks. The ephemeris data can be transmitted much more
quickly from the cell phone than the 50 baud from the satellite. I have heard of
fixes in 5 seconds, certainly 7 is reasonable. For example, you leave a building
where all GPS signals are blocked and with 2 or 3 seconds it has a fix since it
already has the setup data.
Dale described it well but of course I'll add a few notes. :-) |
In our test application, we reset the GPS chip before a test and
trigger the fix with a software meesage. The GPS chip is in low power
mode until the software message is sent. Since we developed the
hardware and software, we can control these things pretty tightly.
In the world of cell phones, keeping the GPS powered all the time is a
battery killer (when did you encounter a GPS with 300 hours of life on
1,500 mAH!). So if it works for the application, you keep the GPS in
low power mode until you really need it. With a quick time to a fix,
you can keep it powered down even longer.
Actually the ephemeris data is available on the internet and the phone
just brings up its data connection, connects to the server across the
internet and grabs the data. The slow part is getting the network
connected. My observations were on a GSM network using GPRS. A phone
with a newer 3G network that is always "on" the data network might
drop the time a bit. (Although I wonder if we start hitting another
limit?)
The other winner in this is having an approximate location and
time. Both help the GPS do its job quicker.
The other big benefit to AGPS is low signal performance. Since you
don't have to decode a long message from the satelites (ephemeris),
there's a greater chance you can get a fix in a lower signal
environment. I think it equates to about 5 or 7 dB which in radio is
pretty huge. In a cell phone, the lousy antenna used for the GPS (to
make them "pretty") more than makes up for this gain!
The Open Mobile Alliance is the group making all of this standard
across the industry. You can find more information with them. |
|
| Back to top |
|
|
|
|
|
|
FAQ
Memberlist
Register
Profile
Log in to check your private messages
Log in