Re: [gpsd-users] Processing RTCM messages

[Michele Pangrazzi]

First, thank you for the precise answer.

As you understood, I initially wanted to use the webserver as a place to 
process RTCM messages, mainly because the GPS device on the Android smartphone 
doesn’t support RTCM and can’t receive DGPS data. The only thing that does via 
native API is emitting NMEA sentences.

Is there a way to obtain required pseudoranges from NMEA sentences, or they are 
meant to stay on the GPS device?

Thank you again,
Michele


> On 09 Feb 2016, at 14:37, Greg Troxel <address@hidden> wrote:
> 
> 
> Michele Pangrazzi <address@hidden> writes:
> 
>> 1. First step
> 
>> Android device (with GPS) —————> NMEA over HTTP POST ——————> Webserver
> 
>> The Android device sends a NMEA string over HTTP POST API to web server, for 
>> example:
>> $GPGGA,150505.00,4608.072256,N,01105.977973,E,1,10,1.2,247.7,M,47.0,M,,*60”
> 
> What's the point of the webserver?   Why can't the device itself just
> talk to the caster?
> 
>> 2. Second step
>> 
>> Webserver —————> NMEA over NTRIP ——————> Local NTRIP caster 
>> (194.105.50.232:2101)
>> 
>> The webserver sends the NMEA string obtained at first step using NTRIP
>> protocol to a local caster. I’ve implemented a simple NTRIP protocol
>> communication using Node.js.
> 
> This may well be what's required, but sending your position to the
> Caster is irregular from an OPSEC* viewpoint.
> 
> * https://en.wikipedia.org/wiki/Operations_security
> 
> In all seriousness, you may want to truncate the coordinates to a tenth
> of a degree or less.  But your caster may be interpolating.
> 
>> 3. Third step
>> 
>> Local caster (194.105.50.232:2101)  —————> RTCM over NTRIP ——————> Webserver
>> 
>> Basically, I receive first a 1013 message (system parameters), then a
>> lot of 1004 / 1012 messages (GPS / GLONASS observations). The NTRIP
>> caster is a Leica
>> SpiderWeb<http://leica-geosystems.com/products/gnss-reference-networks/software/leica-spiderweb>.
> 
> That all sounds ok (modulo OPSEC concerns), and the unclear rationale
> for the intermediate server.
> 
>> I was wondering if there is a way to actually compute the correction
>> on server-side starting from those messages (and the NMEA string
>> obtained at first step). If yes, the webserver could send to the
>> Android device a more precise position, maybe in lat/lon format (it
>> would be great).
> 
> Basically, no.  One needs to apply the calculations to the raw
> pseudoranges, and those stay internal to the smartphone receiver.  This
> first page is much too high level to lead to understanding, but they
> provide a starting point:
> 
>  https://en.wikipedia.org/wiki/Differential_GPS
>  https://en.wikipedia.org/wiki/Pseudorange
> 
> However, you may be able to carry RTCM to the device and inject it to
> the receiver on the phone.  Traditional GPS receivers with serial ports
> like theGarmin GPS V, could accept RTCM over the serial port, such as
> one might receive from a LF DGPS transmitter.  I am unclear on whether
> the GPS chip in phones is wired to do that, and whether there are
> interfaces in the OS.