Transmit GPS data to website

[tracking]

Hello 

I would like to transmit gps data to my website in addition to nrf cloud, there is a example or documentation about this that I can take of reference? I am working with nRF 9160 DK 

thanks in advance.

Kind Regards

Julio

[Håkon Alseth]

Hi,

 

The asset_tracker shows how to connect to a cloud based system, and send certain sensors + gps data to the cloud.

If you want a simpler solution, like a tcp/udp connection directly to your own server, I would recommend that you look at the "gps" sample, and add your own implementation around this.

 

Note that there's a restriction towards getting GPS and LTE concurrent, which is that you have to be in eDRX or PSM mode in order to use the GPS. A recommended use-case is to use the AT CSCON event to signal when you can or cannot use the GPS.

 

Kind regards,

Håkon

[paul_tanner]

Thx Håkon,

I tried setting CSCON=0 when initialising the GPS.

(After looking at the code in asset_tracker I also saw it use lte_lc_psm_req(true))

If I then try to initialise a UDP connection lte_lc_init_and_connect_async(lte_handler) fails.

(Initialisation code from the UDP example)

I'm not clear on the sequence I need to initialise both GPS and UDP.

further...

I wanted to see why lte_lc_init_and_connect_async is failing but this seems to be part of the distribution that is shipped in binary.

Specifically, the error is:

modem_configure()
I: Using legacy LTE PCO mode...
I: PDP Context: AT+CGDCONT=0,"IP","lpwa.vodafone.iot"
E: Could not send AT command, error: -8

Which is odd because when initialising the GPS we already sent 4 AT commands without error.

This part of UDP initialisation works fine if I comment out the GPS initialisation 

Thx.

[paul_tanner]

Many thx Håkon

I guess this means that if testing without a connection it will wait indefinitely - if you let it.

That should not be a problem as you can't send anything until you get a signal.

I'll try this when I'm in range of a cell tower.

Regards, Paul

[Håkon Alseth]

You can evaluate the gps without using LTE, by running the "gps" example.

This doesn't require LTE, but does require that you're outside (or the GPS-antenna shall be outside) while testing.

 

Kind regards,

Håkon

[paul_tanner]

Thx Håkon

Yes.  The first thing I did was to test both GPS and UDP examples.

Only then did I try to put them together to make a minimal tracker using UDP and without the need for all the other good stuff in your asset_tracker app.

So I now wait for udp Idle.

However, from the code in the gps sample I then get an error from

```

at_cmd_write(at_commands[i])

```

AT_XSYSTEMMODE,  (fails on this one)
AT_MAGPIO,
AT_COEX0,
AT_ACTIVATE_GPS

so GPS is not activated.

XSYSTEMMODE is probably redundant but why would it throw an error?

Regards.

[Håkon Alseth]

Hi,

 

paul_tanner said:

XSYSTEMMODE is probably redundant but why would it throw an error?

Its highly likely because the modem is already enabled (AT+CFUN=1) at this point. Please check that "modem_configure()" isn't called, and that "CONFIG_LTE_AUTO_INIT_AND_CONNECT=n" is set in your project.

 

Kind regards,

Håkon

[paul_tanner]

Thx Håkon

modem_configure() *was* being called; CONFIG_LTE_AUTO_INIT_AND_CONNECT=n is set.

So I removed modem_configure().  The result was that we never get the LTE_LC_EVT_RRC_UPDATE (idle) event so it just hangs.

Is something else needed instead of modem_configure()? (per the udp example)

Presumably the lack of modem_configure would also prevent sending of data via udp?

Regards.

[Håkon Alseth]

Hi Paul,

 

My strongest apologies! This comment from me is written a bit too fast (ie: its wrong..)

Håkon Alseth said:

Please check that "modem_configure()" isn't called

You should check that modem_configure hasn't been called yet, before setting system permissions, like XSYSTEMMODE.

So, issue your AT-commands, then call modem_configure() to connect to the LTE network.

 

Kind regards,

Håkon 

[paul_tanner]

Thx Håkon 

I tried that.  It gets to modem_configure() - same code as in UDP sample - and we then fail with:

```

Could not send AT command error -8

```

This is happening inside lte_lc_init_and_connect_async() for which AFAIK we do not have the source.

Regards. 

[Håkon Alseth]

Hi,

 

paul_tanner said:

This is happening inside lte_lc_init_and_connect_async() for which AFAIK we do not have the source.

That function is inside lte_lc.c: https://github.com/nrfconnect/sdk-nrf/blob/master/lib/lte_link_control/lte_lc.c#L744

Do you know which AT command is failing? Does any of them work?

 

Kind regards,

Håkon

[paul_tanner]

Thx Håkon 

OK. modem_configure() is trying to set the AT%XSYSTEMMODE - which is *already* correctly set.

(I have CONFIG_LTE_NETWORK_MODE_NBIOT_GPS=y in prj.conf)

BTW. To find this out I had to edit lte_lc.c whereas I would have preferred to turn on debugging.  I tried adding:

  #define NRF_AT_CMD_LOG_LEVEL NRF_LOG_LEVEL_DEBUG

in main.c but that had no effect.

What next?

Regards.

[paul_tanner]

So I made that error in modem_configure() non-fatal (a hack in ltc_lc.c) and it now does GPS and transmits the result via UDP.  This is almost a solution apart from the fact that I hacked ltc_lc.c.  I'd like a way to do this without that hack if possible to make it more upgrade-proof.

Also, if I want to restrict the log a bit i'd like to find out where CONFIG_LTE_LINK_CONTROL_LOG_LEVEL is defined.

Thx for any suggestions.

Regards