forum.pistar.uk

ok. thanks.

this works. I can actually see GPS data on /dev/ttyUSB0 when I sym link it to rfcomm0
and now I can see the port in mobile GPS configuration of Pi-Star UI.
How can I verify that this location is used ?
Actually where do we see the location ?

And when does pi-star update location ? each time we send data/voice on DMR ?

You are beyond my realm now... And maybe into an area that will be short lived... See viewtopic.php?t=3661

MW0MWZ wrote:MobileGPS is going away shortly in favour of gpsd, I need to sort out how I will handle that in the dashboard / config, it's on my to-do list.

(Though it is still there some 9 months later)

viewtopic.php?t=2983

NQ4T wrote:When I last looked at MobileGPS stuff a couple years ago...it was not geared toward real-time stuff. It seemed like it was more for getting the GPS information the hotspot itself reports to networks. It didn't look like it had a very fast poll time. I did dig around using what was already in there before trying to bolt things on

Upstream -- which may mean it isn't yet in the Pi-Star release (though that was summer of 2020) https://github.com/g4klx/MMDVMHost/comm ... 0d89a8d9bc
MMDVMHOST.cpp implies that GPSD is a compile-time option (so again, may not be active in Pi-Star)
Makefile shows the default build is without GPSD

That said, /etc/mmdvmhost on my Pi-Star system still has the config section for Mobile GPS, so either that is being ignored, or MMDVMHost is older than the GPSD change. I don't really know where to look deeper. Pi-Star log shows: which is a year after the upstream changes. Pi-Star release notes state that Mobile GPS was added 20-May-2020 **Version 4.1.2**, haven't found release notes for 4.1.5/4.1.6.

Added the MobileGPS service, this is a service for using USB GPS dongles to update the position information in MMDVMHost and YSFGateway. Support in ircDDBGateway will follow soon.

Also -- GPSD itself is a separate system service that reads GPS devices, formats position information, and makes it available via socket I/O, not serial port.

so, what I understand is:

1. mobile gps is not supported but still there.
2. mobile gps might be providing only hotspot location once when it is UP.
3. mobile gps is not providing real time location
4. pi-star developer planned to use gpsd , but did not
5. pi-star developer might be moving to gpsd , but no sign of it yet.
6. MMDVMHOST has already implemented gpsd (?) but that version is not used in Pi-star

well, if I sum it up:
- nobody really cares about realtime location on pi-star (including the developer)
- there is no way to provide realtime location from pi-star as of now
- we really don't know what the current "mobile gps" configuration does

Do I understand correct ?

And, there is an aprs configuration on Pi-star, what does it do ?

TB2IAK wrote: ↑Sat Feb 26, 2022 2:48 pm so, what I understand is:

1. mobile gps is not supported but still there.
2. mobile gps might be providing only hotspot location once when it is UP.
3. mobile gps is not providing real time location

Since I don't "do" GITHUB, I don't know how to extract MMDVMHOST source code to find an older version with Mobile GPS still implemented (the differences for the current MMDVMHOST development version has added a file for GPSD, and deleted the file that handled Mobile GPS). As a result, I can not make any statements as to what the code actually did for Mobile GPS.

4. pi-star developer planned to use gpsd , but did not
5. pi-star developer might be moving to gpsd , but no sign of it yet.
6. MMDVMHOST has already implemented gpsd (?) but that version is not used in Pi-star

The "pi-star developer" is dependent upon what he gets from the developers of ircddbgateway, DMR gateway, MMDVMHOST, etc.

When upstream MMDVMHOST changed to gpsd, it meant that the version would not work with current Pi-Star (Pi-Star is just the packaging of all those independent parts, along with the code for the "dashboard"). Past history has indicated that it takes over a year for an MMDVMHOST change to migrated to the Pi-Star dashboard/config -- and that was when the developer was "full time"! Have you read viewtopic.php?t=4061

well, if I sum it up:
- nobody really cares about realtime location on pi-star (including the developer)
- there is no way to provide realtime location from pi-star as of now
- we really don't know what the current "mobile gps" configuration does

Pretty much -- it is where one's radio is that the location becomes meaningful, not the location of the hotspot (or repeater -- since some variations of Pi-Star and boards are used to drive radio pairs to make a true digital repeater). D-STAR radios with on-board GPS can be configured to embed position information (whether that information makes it through the reflector system is a different matter -- though there are something like two standards for how to embed position, and fewer radios can decode it on receive; it is also recommended that one NOT configure the radio to send position on PTT when doing voice as it may just clutter the servers). Since there is no configuration for beacon interval or smart beaconing (a scheme which determines when to send a position based upon speed and turns) Mobile GPS either is a one-time operation, or has some obnoxiously long interval -- otherwise the providers might complain about getting flooded with niggling position reports (have you ever mapped how much wander a consumer GPS reports as the satellites change position -- especially if one might be getting multi-path interference..

The idea of using a "low-power" HT (many of which don't go below 1W) inside a vehicle to reach an even lower-power hotspot, which then forwards the data using WiFi frequencies to a cell-phone configured as a WiFi access point so it can transmit the data on a third frequency, all while in motion, just boggles me... Has anyone ever considered what the RF exposure must be like inside that vehicle with three RF sources creating mixing products, etc.

Fixed locations, like at a camp-site, I can understand... Leave the hotspot and cell-phone in the vehicle, while using the HT(s) while moving around camp (I think my hotspot has at least a 500 ft range, which is basically to both ends of my street, and thats from inside a building with a steel roof and aluminum siding).

You are perfectly free to download the MMDVMHost source files from GITHUB, especially if you can figure out how to revert back to the change that last had Mobile GPS, and attempt to figure out what it is doing.

Do I understand correct ?

And, there is an aprs configuration on Pi-star, what does it do ?

Specifies which (digital) APRS server gets sent location information -- likely on boot-up.

Based upon my investigations, the only confirmed way to report dynamically changing position information requires one to create a shell script (or maybe a Python program doing the same stuff) which grabs NMEA records from some GPS device, parses the information into an APRS message, and sends that APRS message to a known server. Said server can be extracted from the config files -- so changing in Pi-Star config would propagate to the script at some point. This was described in viewtopic.php?t=1724 although the OP was mostly using two Raspberry-Pi systems: one running Pi-Star software, and the other using GPS to create an NTP (time server) system -- the script grabbed GPS data from the time server, and parsed reflector/talkgroup information from Pi-Star (I didn't dig deep enough to see if it was scraping the dashboard or using something deeper).

ok thanks for the great explanation.
I took your time with "currently" useless gps information.
maybe in future, some developer decides to implement a method to send realtime gps information.
I saw the script way. In fact, I can of course write my own scirpt to post gps data to aprs servers (learning the required format)
but that is a long way and really I don't need it.
I was just curious why we have gps configuration on Pi-Star, and if it had worked I could start using it.

thanks.

TB2IAK wrote: ↑Sun Feb 27, 2022 4:27 pm ok thanks for the great explanation.
I took your time with "currently" useless gps information.
maybe in future, some developer decides to implement a method to send realtime gps information.
I saw the script way. In fact, I can of course write my own scirpt to post gps data to aprs servers (learning the required format)
but that is a long way and really I don't need it.
I was just curious why we have gps configuration on Pi-Star, and if it had worked I could start using it.

The primary purpose appears to be to allow one to take a hotspot to some remote location for /fixed/ operation, and not have to manually edit the lat/long fields in the config pages. Get to your campsite or whatever, power up the hotspot with GPS dongle, and it should send the campsite location to the server (if I understood some of the GPSD-based code, it may delays 1000msec [1sec] between checking for a "fix" from the GPS system, and will repeat that up to 60 times [1min total]; I'm guessing MobileGPS did something similar for timing).

Once you attempt dynamic (in motion) updates of location you run into the problem of: how often do you update without spamming the server with position reports.

For my ancient Yaesu VX-8DR, vehicle smart beaconing is
<5MPH beacon every 30 minutes
>70MPH beacon every 2 minutes
beacon if the heading changes by 28 degrees -- but no more than one beacon every 30 seconds while turning (with some tweaking for speed)

Timing is interpolated between those end-point speeds. Essentially, for a vehicle, one is looking at a beacon every 2.5 miles of straight line travel, maxed at 70MPH/2min intervals

{bicycle is <3MPH/>30MPH, timing is still the same; walking is <2MPH/>12MPH}