Re: AGPS - protocol specs?
On Fri Mar 9 13:46:19 CET 2007, Al Johnson wrote: Here's a civil device using RTK that determines the position to centimeter accuracy within a few seconds at ranges up to 50 km from a reference station Sounds like an interesting challenge. At least rtklib 2.4.0 was unable to resolve the integer ambiguity problem (?) when I extended the baseline to about 30 km. Would be nice to know if this is due to my cheap antenna, finnish weather, a problem in rtklib or something completely different. (Yes, I did turn off all radio transmitters during the tests and ran from battery power only). ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
Hello. Nils Faerber napisał(a): There is all sorts of wacky stuff - for example, peer-peer DGPS that can be done, where all stationary neos on charge with a GPS signal and a free internet connection contribute to a global ionospheric model. Then any Neo can connect to this model, download 200 bytes or so, and get +-0.3m (or better) position for a short while. Exactly. Also relative positioning can be made much more precise using the raw data (AFAIK in the range of cm not m). I asked my colleagues who are GPS devices specialists and they said that this is all wrong. Military devices can do magic but civil not. They said that with civil GPS receiver you can get accuracy up to about 5 - 10 meters. For such precision it is required to be seen at least 8 satellites, a clear sky and good magnetic and ionosphere conditions (also solar magnetic field's change is important). To get *any* GPS reading device must see 3 satellites but you cannot be sure the position. It can vary even up to hundreds of meters! Of course you can get very accurate reading but device must see many satellites and stand up still for many hours! Then the mean value is computed and can be assumed that GPS position reading is accurate to centimeters... You cannot depend on GPS reading to count quite accurate velocity, acceleration and accurate position. Sorry. Regards, -- *Bartlomiej Zdanowski* Programmer Product Research Department AutoGuard Insurance Ltd. Omulewska 27 street 04-128 Warsaw Poland phone +48 22 611 69 23 www.autoguard.pl http://www.autoguard.pl ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
2007/3/9, Bartlomiej Zdanowski AutoGuard Ltd. [EMAIL PROTECTED]: I asked my colleagues who are GPS devices specialists and they said that this is all wrong. Did they talked about AGPS (http://wiki.openmoko.org/wiki/Hardware:AGPS)? cayco ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
Krzysztof Kajkowski napisał(a): 2007/3/9, Bartlomiej Zdanowski AutoGuard Ltd. [EMAIL PROTECTED]: I asked my colleagues who are GPS devices specialists and they said that this is all wrong. Did they talked about AGPS (http://wiki.openmoko.org/wiki/Hardware:AGPS)? They did about technology itself not about marketing terms (as Wiki says): AGPS is purely a marketing term. Hot start has nothing to do with it. GPS is still very inaccurate. It's very powerful if you focus on task that Object (human, car, plane etc.) that carries GPS receiver does. Lets assume that an Object is a car and you have accurate GPS maps. Car equipped with GPS receiver drives only through road and not surrounding buildings and side-walks that's sure. So now even if reading does not point somewhere on the road software can compute new position (depending on car's speed and direction) to correct red GPS position to put car on the road. Understand this? If person should walk on side-walk, with current area map you can correct GPS inaccurate reading to compute new position on side-walk instead somewhere near on road, in build or in surrounding bushes. I work for company that develops devices for car and person tracking. We do the stuff as written above to correct GPS readings. A car that is parked in front of our office equipped in GPS receiver collects data about it position. Even if is parked in the same place for all day, with a clear sky above (a many satellites visible), raw GPS readings show that it hangs around all day in range of a few tens of meters and sometimes a hundred of meters. And I can assure the we use good GPS receivers. The problem is very complicated. -- *Bartłomiej Zdanowski* Programista Dział Rozwoju Produktów AutoGuard Insurance Sp. z o.o. Sąd Rejonowy dla m.st. Warszawy, XIII Wydział Gospodarczy Krajowego Rejestru Sądowego KRS: 029534 NIP PL1132219747 ul. Omulewska 27 04-128 Warszawa tel. +48 22 611 69 23 www.autoguard.pl http://www.autoguard.pl begin:vcard fn;quoted-printable:Bart=C5=82omiej Zdanowski n;quoted-printable:Zdanowski;Bart=C5=82omiej org;quoted-printable:AutoGuard Insurance Sp. z o.o.;Dzia=C5=82 Rozwoju Produkt=C3=B3w adr:;;ul. Omulewska 27;Warszawa;;04-128;Polska email;internet:[EMAIL PROTECTED] title:Programista AC2 tel;work:022 611 69 23 tel;cell:603 525 105 x-mozilla-html:TRUE url:http://www.autoguard.pl version:2.1 end:vcard ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
Al Johnson napisał(a): On Friday 09 March 2007 08:58, Bartlomiej Zdanowski AutoGuard Ltd. wrote: You cannot depend on GPS reading to count quite accurate velocity, acceleration and accurate position. Sorry. Conventionally this is true, but there are other ways. IIRC RaceLogic use the doppler shift of the GPS signal for speed and acceleration calculations. These can also be used to improve the accuracy of the location. I think they claim it's good enough to compare the line taken by the driver on successive laps of the race circuit. Yes, my GPS-wise colleagues said that measuring difference between two sequential readings it can be possible to compute acceleration because reading error is quite same in both readings. So while global position is not accurate, difference is accurate. Another problem is reading delay so you have not current data (velocity) but data delayed with reading period's time. It's not a problem since human don't travel fast on foot. It can be while trying to compute acceleration of a vehicle - it changes very fast. Regards -- *Bartlomiej Zdanowski* Programmer Product Research Department AutoGuard Insurance Ltd. Omulewska 27 street 04-128 Warsaw Poland phone +48 22 611 69 23 www.autoguard.pl http://www.autoguard.pl ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
On Friday 09 March 2007 09:58, Bartlomiej Zdanowski AutoGuard Ltd. wrote: To get *any* GPS reading device must see 3 satellites but you cannot be Actually, for a single location fix, you need four. With 3 sources you can still have 2 spots on the surface of the Earth that can match up (some GPS receivers being smart and discarding the 2nd solution based on likelihood is a different story). You cannot depend on GPS reading to count quite accurate velocity, acceleration and accurate position. Sorry. Position is a different beast in comparison to velocity and acceleration, here the term accurate and precise start differing quite a bit - you can have a precise reading (let's say 1 cm) and, as strange as it may seam, that does NOT mean that your location is +-1cm from what you read on the display (you can try this with several GPS-es in one place - they will have differeces much higher than their precision). ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
Ian Stirling napisał(a): Bartlomiej Zdanowski AutoGuard Ltd. wrote: Hello. Nils Faerber napisał(a): There is all sorts of wacky stuff - for example, peer-peer DGPS that can be done, where all stationary neos on charge with a GPS signal and a free internet connection contribute to a global ionospheric model. Then any Neo can connect to this model, download 200 bytes or so, and get +-0.3m (or better) position for a short while. Exactly. Also relative positioning can be made much more precise using the raw data (AFAIK in the range of cm not m). I asked my colleagues who are GPS devices specialists and they said that this is all wrong. Military devices can do magic but civil not. They said that with civil GPS receiver you can get accuracy up to about 5 - 10 meters. For such precision it is required to be seen at least 8 satellites, a clear sky and good magnetic and ionosphere conditions (also solar magnetic field's change is important). It's a little more complex than that, and not quite as bad. http://www.mauve.demon.co.uk/gps-average.gif is some data I took a few years back with a garmin GPS12. The circles show radiuses inside which the stated number of points fall, in a 10 second average. For example, 99.99% of points fall within 13.8m. 10s averages are red dots, green 100 second, magenta circles 1 hour, cyan squares 6 hours, and black 24h. Try it while walking on streets between buildings :( Another problem can be when we put Neo out of pocket. I don't know if it's GPC received would need some extra time to compute pos. If a number of factors in the hardware all come together (that details are not public on) position under 10cm may be possible after downloading the few hundred bytes of correction information. (It may be possible to optimise this down a lot) So *military* devices does it. But not civil. -- *Bartlomiej Zdanowski* Programmer Product Research Department AutoGuard Insurance Ltd. Omulewska 27 street 04-128 Warsaw Poland phone +48 22 611 69 23 www.autoguard.pl http://www.autoguard.pl ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
Bartlomiej Zdanowski AutoGuard Ltd. wrote: Ian Stirling napisał(a): Bartlomiej Zdanowski AutoGuard Ltd. wrote: Hello. Nils Faerber napisał(a): There is all sorts of wacky stuff - for example, peer-peer DGPS that can be done, where all stationary neos on charge with a GPS signal and a free internet connection contribute to a global ionospheric model. Then any Neo can connect to this model, download 200 bytes or so, and get +-0.3m (or better) position for a short while. Exactly. Also relative positioning can be made much more precise using the raw data (AFAIK in the range of cm not m). I asked my colleagues who are GPS devices specialists and they said that this is all wrong. Military devices can do magic but civil not. They said that with civil GPS receiver you can get accuracy up to about 5 - 10 meters. For such precision it is required to be seen at least 8 satellites, a clear sky and good magnetic and ionosphere conditions (also solar magnetic field's change is important). It's a little more complex than that, and not quite as bad. http://www.mauve.demon.co.uk/gps-average.gif is some data I took a few years back with a garmin GPS12. The circles show radiuses inside which the stated number of points fall, in a 10 second average. For example, 99.99% of points fall within 13.8m. 10s averages are red dots, green 100 second, magenta circles 1 hour, cyan squares 6 hours, and black 24h. Try it while walking on streets between buildings :( Another problem can be when we put Neo out of pocket. I don't know if it's GPC received would need some extra time to compute pos. Nope - the chipset gets a 5m position in one second according to the brief datasheet. If a number of factors in the hardware all come together (that details are not public on) position under 10cm may be possible after downloading the few hundred bytes of correction information. (It may be possible to optimise this down a lot) The DGPS corrections - per satellite error - helps you a little in urban canyons, as if you can see several satellites, and their errors check with what you expect, you can throw out reflections that don't correlate with expected errors. (well, not throw out, but assume to be a reflection) So *military* devices does it. But not civil. No, I'm _not_ talking about the more precise 'P' code - simply accuracy of phase tracking loops, and how 'tunable' the GPS hardware is for the C/A code. DGPS can give results down to centimetre level. ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
On Friday 09 March 2007 11:43, Bartlomiej Zdanowski AutoGuard Ltd. wrote: Ian Stirling napisał(a): Bartlomiej Zdanowski AutoGuard Ltd. wrote: Hello. Nils Faerber napisał(a): There is all sorts of wacky stuff - for example, peer-peer DGPS that can be done, where all stationary neos on charge with a GPS signal and a free internet connection contribute to a global ionospheric model. Then any Neo can connect to this model, download 200 bytes or so, and get +-0.3m (or better) position for a short while. Exactly. Also relative positioning can be made much more precise using the raw data (AFAIK in the range of cm not m). I asked my colleagues who are GPS devices specialists and they said that this is all wrong. Military devices can do magic but civil not. They said that with civil GPS receiver you can get accuracy up to about 5 - 10 meters. For such precision it is required to be seen at least 8 satellites, a clear sky and good magnetic and ionosphere conditions (also solar magnetic field's change is important). It's a little more complex than that, and not quite as bad. http://www.mauve.demon.co.uk/gps-average.gif is some data I took a few years back with a garmin GPS12. The circles show radiuses inside which the stated number of points fall, in a 10 second average. For example, 99.99% of points fall within 13.8m. 10s averages are red dots, green 100 second, magenta circles 1 hour, cyan squares 6 hours, and black 24h. Try it while walking on streets between buildings :( Another problem can be when we put Neo out of pocket. I don't know if it's GPC received would need some extra time to compute pos. If a number of factors in the hardware all come together (that details are not public on) position under 10cm may be possible after downloading the few hundred bytes of correction information. (It may be possible to optimise this down a lot) So *military* devices does it. But not civil. This is for *civil* devices. http://pro.magellangps.com/en/products/aboutgps/dgps.asp http://pro.magellangps.com/en/products/aboutgps/rtk.asp http://en.wikipedia.org/wiki/Real_Time_Kinematic Here's a civil device using RTK that determines the position to centimeter accuracy within a few seconds at ranges up to 50 km from a reference station http://www.leica-geosystems.com/uk/en/lgs_8276.htm In the UK the Ordnance Survey is making its data available for commercial RTK networks. http://www.ordnancesurvey.co.uk/oswebsite/gps/commercialservices/index.html ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
Bartłomiej Zdanowski DRP AC2 [EMAIL PROTECTED] writes: I work for company that develops devices for car and person tracking. We do the stuff as written above to correct GPS readings. A car that is parked in front of our office equipped in GPS receiver collects data about it position. Even if is parked in the same place for all day, with a clear sky above (a many satellites visible), raw GPS readings show that it hangs around all day in range of a few tens of meters and sometimes a hundred of meters. And I can assure the we use good GPS receivers. I wonder if you'd do better with a slightly better GPS or better GPS antenna. With a good consumer GPS and an unobstructed sky view all around you shouldn't be averaging a 10 meter error. Even a 10 year old, gps designed during the days of Seleective Availability did ~2.5 meters in a 24hr test here. I would expect the modern units to be even tighter. http://www.wsrcc.com/wolfgang/gps/accuracy.html In practice when overlaying gps tracklogs with geo-referenced overhead imagery I regularly see good enough tracking to have a very good guess as to which lane the car was driving in. Look at the exit ramp screenshot (second from the top) to see what I mean. http://www.gnomad-mapping.com/screenshots-new/ And yes, I'm going to try to port this mapping program to openmoko. The real fun starts when everyone you know has one and you can see where they are on the map in real-time. -wolfgang -- Wolfgang S. Rupprechthttp://www.wsrcc.com/wolfgang/ ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
On Fri, 2007-03-09 at 12:46 +, Al Johnson wrote: ... So *military* devices does it. But not civil. This is for *civil* devices. http://pro.magellangps.com/en/products/aboutgps/dgps.asp http://pro.magellangps.com/en/products/aboutgps/rtk.asp http://en.wikipedia.org/wiki/Real_Time_Kinematic Here's a civil device using RTK that determines the position to centimeter accuracy within a few seconds at ranges up to 50 km from a reference station http://www.leica-geosystems.com/uk/en/lgs_8276.htm In the UK the Ordnance Survey is making its data available for commercial RTK networks. http://www.ordnancesurvey.co.uk/oswebsite/gps/commercialservices/index.html Very interesting thread. Took the liberty of adding some ideas entitled Cooperative Differential GPS related to this to the Summer_of_code Wiki page: http://wiki.openmoko.org/wiki/Summer_of_code Please edit as appropriate to expand/correct. Phil ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
Philip Ray Schaffner wrote: On Fri, 2007-03-09 at 12:46 +, Al Johnson wrote: ... So *military* devices does it. But not civil. This is for *civil* devices. http://pro.magellangps.com/en/products/aboutgps/dgps.asp http://pro.magellangps.com/en/products/aboutgps/rtk.asp http://en.wikipedia.org/wiki/Real_Time_Kinematic Here's a civil device using RTK that determines the position to centimeter accuracy within a few seconds at ranges up to 50 km from a reference station http://www.leica-geosystems.com/uk/en/lgs_8276.htm In the UK the Ordnance Survey is making its data available for commercial RTK networks. http://www.ordnancesurvey.co.uk/oswebsite/gps/commercialservices/index.html Very interesting thread. Took the liberty of adding some ideas entitled *Cooperative Differential GPS* related to this to the Summer_of_code http://wiki.openmoko.org/wiki/Summer_of_code Wiki page: http://wiki.openmoko.org/wiki/Summer_of_code I can't get to that page for some reason - and am going to sleep. I expounded on this in some length at the bottom of Hardware:GPSD - if you haven't seen that, it may be of use. ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
Ian Stirling schrieb: Nils Faerber wrote: Hi! I am interested in doing a little research on the AGPS part. As far as I understood the used Hammerhead chip will dump out the more or less raw GPS data into userland which will then need to be post-processed in order to get the usual NMEA or whatever messages. So, can OpenMoko provide at least somemore information about that chip and especially its output? No, the chip is under NDA. Darn. I do not care aout most of the chip, just the output format ;) http://wiki.openmoko.org/wiki/Hardware:AGPS has the current state of play. Basically, the satellite format is known, and not especially hard to decode especially if we also get NMEA (or whatever) output out of the supplied daemon at the same time to decode it. You mean some kind of reverse engeneering trying to match the NMEA output to the raw input? There is all sorts of wacky stuff - for example, peer-peer DGPS that can be done, where all stationary neos on charge with a GPS signal and a free internet connection contribute to a global ionospheric model. Then any Neo can connect to this model, download 200 bytes or so, and get +-0.3m (or better) position for a short while. Exactly. Also relative positioning can be made much more precise using the raw data (AFAIK in the range of cm not m). Cheers nils faerber -- kernel concepts GbRTel: +49-271-771091-12 Sieghuetter Hauptweg 48Fax: +49-271-771091-19 D-57072 Siegen Mob: +49-176-21024535 -- ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
Nils Faerber wrote: Ian Stirling schrieb: Nils Faerber wrote: Hi! I am interested in doing a little research on the AGPS part. As far as I understood the used Hammerhead chip will dump out the more or less raw GPS data into userland which will then need to be post-processed in order to get the usual NMEA or whatever messages. So, can OpenMoko provide at least somemore information about that chip and especially its output? No, the chip is under NDA. Darn. I do not care aout most of the chip, just the output format ;) There probably isn't a clear line. http://wiki.openmoko.org/wiki/Hardware:AGPS has the current state of play. Basically, the satellite format is known, and not especially hard to decode especially if we also get NMEA (or whatever) output out of the supplied daemon at the same time to decode it. You mean some kind of reverse engeneering trying to match the NMEA output to the raw input? The gpsd will output assorted parameters, including current position, ... You can take this - or even output from a nearby (1m) GPS, and the bitstream input and output to the GPS chip, and try to work out what the chip sends out. ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
Nils Faerber skrev: Exactly. Also relative positioning can be made much more precise using the raw data (AFAIK in the range of cm not m). Wow! If this is true, speed and direction can be fairly accurate calculated even at low speed. This makes orienting a map with the moving direction towards the top of the screen useful when walking... a virtual compass! Also short distant measuring (a few meters) can be useful! Wish You all luck! PS This would somewhat compensate for the lack of accelerometer and gyro... but would on the other hand make them more useful as You at even low speed can calculate the 'start' vector and you would get more interesting stuff out of the 'realtime' data provided. ie, pointing at buildings etc with gesture should really work (if You move, but slow walking would be enough)! DS /LaH ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
Message: 10 Date: Thu, 08 Mar 2007 13:19:40 + From: Ian Stirling [EMAIL PROTECTED] The gpsd will output assorted parameters, including current position, ... You can take this - or even output from a nearby (1m) GPS, and the bitstream input and output to the GPS chip, and try to work out what the chip sends out. I don't understand what is meant by a nearby (1m) GPS. How does the device communicate with a nearby GPS? Marty ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
Still confused on the term nearby GPS Marty Ian Stirling wrote: Martin Lefkowitz wrote: Message: 10 Date: Thu, 08 Mar 2007 13:19:40 + From: Ian Stirling [EMAIL PROTECTED] The gpsd will output assorted parameters, including current position, ... You can take this - or even output from a nearby (1m) GPS, and the bitstream input and output to the GPS chip, and try to work out what the chip sends out. I don't understand what is meant by a nearby (1m) GPS. How does the device communicate with a nearby GPS? It doesn't - at all. However, in the absence of a GPS daemon that understands the chip output, if the chip could just be turned on, the nearby GPS could be used to determine (almost exactly) what signals are going into the chip, which makes decoding the protocol easier. ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
Martin Lefkowitz wrote: Still confused on the term nearby GPS A completely separate GPS unit, that is nearby, close in location, with almost the same position, ... Its only purpose is to measure the GPS coordinates - lat, long, time, from which can be derived the satellite signals sent to the neo. ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
On Thursday 08 March 2007 20:18:50 Ian Stirling wrote: A completely separate GPS unit, that is nearby, close in location, with almost the same position, ... Its only purpose is to measure the GPS coordinates - lat, long, time, from which can be derived the satellite signals sent to the neo. Seems a lot easier to simply use the ones coming from the closed source module no? pgpZliSSR3u8C.pgp Description: PGP signature ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
Gabriel Ambuehl wrote: On Thursday 08 March 2007 20:18:50 Ian Stirling wrote: A completely separate GPS unit, that is nearby, close in location, with almost the same position, ... Its only purpose is to measure the GPS coordinates - lat, long, time, from which can be derived the satellite signals sent to the neo. Seems a lot easier to simply use the ones coming from the closed source module no? Dramatically, if you can get it. The closed source module is not likely to appear for a month or so AIUI though. ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
OK, now I was the perpetrator of imprecise language. How do the two neo's communicate with each other -- the one that can see the GPS signal and the one that can't? Marty Ian Stirling wrote: Martin Lefkowitz wrote: Still confused on the term nearby GPS A completely separate GPS unit, that is nearby, close in location, with almost the same position, ... Its only purpose is to measure the GPS coordinates - lat, long, time, from which can be derived the satellite signals sent to the neo. ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: AGPS - protocol specs?
Martin Lefkowitz writes: OK, now I was the perpetrator of imprecise language. How do the two neo's communicate with each other -- the one that can see the GPS signal and the one that can't? Not two NEOs -- one NEO and one GPS receiver that's used to get known data. Try to correlate the output from the GPS chip on the NEO with the output from the GPS receiver to decode the protocol used by the chip on the NEO. ___ OpenMoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community