This object is in archive!
Beginner question about NMEA data and postprocessing
Answered
Hello, thanks for this great app!
I have a very basic question about NMEA data.
Some time ago a GPS expert told me it is possible to post-process the data received by a GPS unit in order to produce a more accurate result for recorded tracks and waypoints: I cannot remember the details but I think he mentioned RINEX, WAAS and EGNOS.
He also told me he knew of no standard outdoor GPS unit being able to collect the data necessary to do that (he mentioned some Trimble units equipped to do that, but he also told me they are expensive).
My question is: can NMEA data collected by Locus be used to perform such activity?
If this is possible, how can it be done? Is some kind of howto guide available?
Thanks a lot!
The same question
Good day Hilbert,
Your friend was correct. With complete recorded NMEA data is able to filter location that should be theoretically little bit more or less precise then values that are directly visible on application itself. Most known method of filtering is called "Kalman filter".
In Locus settings > GPS & Location, you may enable recording of received NMEA data, so Locus store everything to txt files in Locus/data/nmea directory for later post-processing.
Anyway method to do this filtering correctly is not so simple and if I remember correctly, best is combination of received NMEA data with some external values like mentioned Egnos data. I'm sure there will be some applications for desktop that do this optimizations, but because I never needed them, I cannot help here. I'm sure that your friend may give you more precise information.
There already exists topic to implement Kalman filtering directly in Locus, but I've not yet found enough time to look on it more precisely. Hope I gave you at least partial answer. If there is something I may help with, let me know.
Good day Hilbert,
Your friend was correct. With complete recorded NMEA data is able to filter location that should be theoretically little bit more or less precise then values that are directly visible on application itself. Most known method of filtering is called "Kalman filter".
In Locus settings > GPS & Location, you may enable recording of received NMEA data, so Locus store everything to txt files in Locus/data/nmea directory for later post-processing.
Anyway method to do this filtering correctly is not so simple and if I remember correctly, best is combination of received NMEA data with some external values like mentioned Egnos data. I'm sure there will be some applications for desktop that do this optimizations, but because I never needed them, I cannot help here. I'm sure that your friend may give you more precise information.
There already exists topic to implement Kalman filtering directly in Locus, but I've not yet found enough time to look on it more precisely. Hope I gave you at least partial answer. If there is something I may help with, let me know.
Hello Menion,
thanks a lot for your answer.
Unfortunately the GPS expert I refer to is not a friend of mine; I met him at a conference and we had a chat there, but it didn't occur to me to ask for his email address, also because I thought it was easy to find info about this subject on the Web.
Later I discovered it is not easy at all unless you are an expert yourself: I found some forums where people discuss about pseudorange, full carrier phase data, gps hackering, self made Arduino hw extensions, and so on, but I never found an easy tutorial teaching an average GPS enthusiast the basic concepts and an easy way to perform post-processing.
For some time I gave up, since I realized no common outdoor handheld GPS unit was able to give the key raw material needed to postprocess data, then I found out about Locus ability to save NMEA data, and this immediately makes me interested again.
I know some theory about linear quadratic estimation (which I think is the same as Kalman filter), but I would not know how to match NMEA data with tracks and waypoints gpx data to get a better estimate of their correct value.
As you say, it would be advisable to find some applications for desktop that do this optimizations instead of trying to do it myself from scrath, but I have not found one that is:
- easy to use, i.e. something like: input GPX, NMEA data and the external source of augmentation (like Egnos), and get the output with postprocessed data
- not tied to a single specific brand of HW GPS units
- possibly free or even better open source
Hello Menion,
thanks a lot for your answer.
Unfortunately the GPS expert I refer to is not a friend of mine; I met him at a conference and we had a chat there, but it didn't occur to me to ask for his email address, also because I thought it was easy to find info about this subject on the Web.
Later I discovered it is not easy at all unless you are an expert yourself: I found some forums where people discuss about pseudorange, full carrier phase data, gps hackering, self made Arduino hw extensions, and so on, but I never found an easy tutorial teaching an average GPS enthusiast the basic concepts and an easy way to perform post-processing.
For some time I gave up, since I realized no common outdoor handheld GPS unit was able to give the key raw material needed to postprocess data, then I found out about Locus ability to save NMEA data, and this immediately makes me interested again.
I know some theory about linear quadratic estimation (which I think is the same as Kalman filter), but I would not know how to match NMEA data with tracks and waypoints gpx data to get a better estimate of their correct value.
As you say, it would be advisable to find some applications for desktop that do this optimizations instead of trying to do it myself from scrath, but I have not found one that is:
- easy to use, i.e. something like: input GPX, NMEA data and the external source of augmentation (like Egnos), and get the output with postprocessed data
- not tied to a single specific brand of HW GPS units
- possibly free or even better open source
Hello,
ah understand.
Well, I'm personally unable to help with software itself. You have to find out on your own or maybe someone other here, will have any suggestions. Anyway, just - don't expect miracles here. NMEA data do not contains raw measured time & satellite values that are used for compute of your location. NMEA just contains already computed values of your location. So what you may do, is some filtering, not a lot more.
To get really precise values, it¨s needed to get data in RINEX format - original measured values. But this is not possible on Android device as I know. For this is needed more accurate and mainly more costly devices, usually based on older Windows Mobile software.
Hello,
ah understand.
Well, I'm personally unable to help with software itself. You have to find out on your own or maybe someone other here, will have any suggestions. Anyway, just - don't expect miracles here. NMEA data do not contains raw measured time & satellite values that are used for compute of your location. NMEA just contains already computed values of your location. So what you may do, is some filtering, not a lot more.
To get really precise values, it¨s needed to get data in RINEX format - original measured values. But this is not possible on Android device as I know. For this is needed more accurate and mainly more costly devices, usually based on older Windows Mobile software.
Replies have been locked on this page!