[OSGeo-Discuss] standard for pushingdata from GPS device to a web service

[Mateusz Loskot]

Anton Martchukov wrote:
> Hello List,
> 
> we want to develop a system including a hardware GPS unit
> with some wireless inet access that is able to push position
> data to the specified web service where it may be stored,
> processed and tracked online as needed. 

Hi Anton,

By the way, are you going to make your application or some of its
components an Open Source Software?
I have similar project on the horizon - an Open Source project - so
may be we could join our forces?
I'm just loosely asking.

> My friend has a commercial order on such a device, but since
> I also though about similar approach for needs of our bike
> club, we decided to unite and make the system as much open
> as possible. 

Sounds cool!

> We had no GPS/GIS experience previously, but after some
> research, we found NMEA protocol that may be used for this
> at some way, but it's proprietary and so is not suitable for
> open systems. 

Am I correct you're going to develop your own GPS device?

If you're not and if you'll buy one of available GPS devices,
then you most likely has no option than processing NMEA messages.

> At this point, we only see just simple requirements for the
> protocol:

Please, could you explain what you mean as "protocol" here?
Do you mean protocol for communication between GPS data reader
and GPS receiver device?
Or protocol for data exchange between GPS data processing software?

> 1) GPS unit sends current coordinates and a timestamp to
> webservice each time it was programmed to send this.
> The protocol should not keep any state, so each message can be
> queued and send as soon as internet connectivity is
> available. Since timestamp is included in each message it is
> not a problem to identify messages for a web service.

Is GPS unit supposed to communicate remote services directly?
Or through middle-ware like user's PDA or laptop?

> Our approach now is to develop a new protocol based on SOAP
> over HTTP, since we can utilise HTTP auth and SSL abilities
> this way and we will not need to mess with data representation.
> Writting WSDL with the data we need to pass from the device
> is enough. 

OK, but I'm not sure I understand where the service client will live.
Here, I thin it's good to define what you mean as "GPS unit".
Simple GPS is only a composite of GPS chip + antena + communication
interface, but another can understand GPS unit as Trimble GeoExplorer
or MobileMapper CE devices which are Windows CE-based.

Obviously, running HTTP client on GPS mouse or simple Bluetooth GPS
units may be a challenge, but on powerful devices like from Trimble or
Magellan it shouldn't be any problem.

Talking about a protocol, have you considered to OGC WFS
for this purpose?

I think WFS or WFS-T should be good to start with.
Next, if reasonable, a customized version of WFS could be developed,
like GFS - GPS Feature Service (tm) ;-)

> Surely, we may not see all the meat here and possible
> extensions and bottle necks in our approach. Or maybe there
> is already a standard protocol available and we will just
> need to stick with it and do not reinvent the wheel. 

I'd suggest to define "gps unit" and "protocol" better,
then it will be easier to iterate through your requirements
with better look.

> So, any advices from the list are welcomed, since we wanna
> make a standard and open solution for this. If someone is
> interesting to cooperate, feel free to do it. 

I believe there is a huge interest of such solution here.

FYI, we've discussed Open Source Mobile GIS solutions here, in OSGeo
Community, some time ago.
Also, we set up the Mobile GIS maling list

http://mobile.maptools.org/index.phtml?page=mailinglist.html

and a project (currently, no product(s) is available, but it gathers
interested users):

http://mobile.maptools.org/

with Wiki page where we put some brainstorm:

http://wiki.osgeo.org/index.php/Mobile_Solutions

Note, that the Mobile GIS initiative is not officially lead by any of
OSGeo projects, but in fact, it's very based on our community.
I believe that after we will have some real development started, then we
could try to request for OSGeo incubation process for the project.


Best regards
-- 
Mateusz Loskot
http://mateusz.loskot.net