#17 own calculation of agps-epo-data


first - thx for your great software!

it would be a good feature if the software could calculate the epo-data for the
new MTK-chipset (iBlue 747A+) locally, not only downloading it from a remote server which may be unavailable.
is this possible? does anybody know the file format of the MTK7d.EPO file?
is there a open source algorithm/software to calculate the epo-data on my own computer?
thx, kalle


  • Mario De Weerd

    Mario De Weerd - 2009-05-26

    As far as I coiuld see, the format is not known publicly, but I do know that the smallest unit is 60 bytes. 3 bytes are for the date the data is valid for and one byte to indicate the satellite the data is valid for.
    The other data is likely similar to the data sent by the GPS sats for the current period.
    One still needs to get some recent data from the NASA if one wants to calculate the values correctly.
    I can give you some more references to resources on the web I gathered if you want to try and find a relationship between publicly known values and the values in the 60 byte record.

    By the way, there is also an MTK14.EPO file available the has data for 14 days.

  • K K

    K K - 2009-05-27

    thx, please give me the links and infos that you have about this. maybe i can find some more infos or assist in some other way.

    do you know how exact the data has to be? and how exact could a "pro" prediction be for 14 days - even with recent nasa data? it should be possible to calculate a few hours into the future with the the almanach and satellite data, maybe only for 12 hours or so.
    are these satellite really so unstable? i thought the estimated position error is max a few seconds on all the orbits per day? have you infos about this?

  • Mario De Weerd

    Mario De Weerd - 2009-05-27

    Here are some results of what I looked into recently.
    The information might also need to take into account the oscillations of the earth, etc.
    I do not intend to look into this topic further because after going through the information that you can find below, I concluded that the effort required is not in line with the benefits and the effort that I want to put into it.
    I'll remind you that there are two things to find out:
    1) The data that is stored in the EPO file;
    2) How to calculate this data.

    You might need to know item 2 before being able to find out about 1.

    However, if you want to extend the validity of satellite data stored in the EPO file this is easy to do because it involves only changing the dates. This is something that could be set up in BT747 for you to test. For instance, you could take old EPO data, apply the current date, upload and check the behaviour.

    If I were to design AGPS, I would store the same data as I am receiving from the satellites.
    So I went looking what that was and it is:
    GPS broadcast ephemeris

    which seems to be in RINEX format (http://www.geocities.com/Athens/Acropolis/2442/rinex_tbl.html).

    This seems to be confirmed in (select specification):

    where some of the needed data is described (which can also be found in the RINEX format description).
    That site also is an AGPS provider.

    Past and near future GPS broadcast ephemeris can be found at:

    Apparently it may also be available in RCTM format as mentioned on this page of another provider:

    There is another article that may be of interest, but personally I did not see usefull information in this context:
    http://arstechnica.com/gadgets/news/2009/01/assisted-gps.ars/1 .

    Some other keywords (SUPL) are mentioned here and could lead to more interesting directed searches.


  • K K

    K K - 2009-05-27

    thx for your infos

  • K K

    K K - 2009-05-28

    i just found this source code of decoding the raw gps-ephemeris-data,



    maybe i can identify some of the parameters in the epo-file.
    the raw data is stored in a bit-oriented structure,
    but i hope the epo-data -file is in byte-size-structure already.

  • Mario De Weerd

    Mario De Weerd - 2009-05-28

    Interesting project that your refer to.
    I notice a refernce to a paper document in the comments that you might not want to mis out from.
    You can find it like this:



Log in to post a comment.

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:

No, thanks