Limited GPS coordinate precision: device or FrSky telemetry limitation?

[dylanbeaudette]

Hi everyone,

I've been collecting some interesting flight data via telemetry on my FrSKY transmitter with this interesting GPS module:

FrSky GPS ADV Sensor

The FrSky GPS ADV sensor provides accurate synchronization with GPS satellites and provides telemetry for groundspeed, GPS position, UTC Time and altitude. It works with any Smart port capable receiver but what's new is F.Port/FBUS capability, allowing for even faster telemetry. Dimensions 45 x...

alofthobbies.com

In short, this device interfaces with an SBUS or F.Port compatible receiver and sends back all kinds of interesting "sensor" data back as telemetry. GPS date/time, elevation, speed, and coordinates. For example, speed:

Note that the my time zone offset hasn't been correctly set--I wasn't actually flying at 2am in the morning. Still there is sufficient (apparent) precision to do a lot with these data in terms of understanding flight characteristics--maybe.

Looking at the GPS coordinates, I realize that something isn't right; they appear to be truncated. Note the snapping effect in the following map:

Looking over the raw data, from the transmitter log file, I see that the "GPS" sensor is returning a single string containing latitude and longitude combined:

"38.032500 -120.407600"

For some reason, the last two digits have been set to 0. On the transmitter, in the telemetry display, the GPS coordinates are formatted in "degree minute second" (DMS) notation with 2 decimal places in the seconds. Like this: -120° 33′ 5.36″. This formatting implies 4 decimal places of precision (120.5649) when using decimal degrees (DD).

OK, the telemetry log matches the telemetry screen even though one is formatted as DMS and the other as DD. Two questions

1. does the transmitter have access to the full precision possible with GPS?
2. is the GPS device only capable of sending this level of precision due to telemetry protocol limitations

Locations described by 4 decimal places of DD notation are only capable of resolving roughly 30m grid cells at mid-latitudes, hence the grid-like pattern in the map above.

I'm wondering if a more advanced telemetry processing script could solve the problem. For example: https://github.com/moschotto/OpenTX_GPS_Telemetry

I'll give this a try and report back.

Dylan

 

[LitterBug]

The documentation for the module is vague. I have always used a serial GPS hooked up to a FC or OSD and had better precision.

The manual and firmware can be found here: https://www.frsky-rc.com/gps/

 

[dylanbeaudette]

The documentation for the module is vague. I have always used a serial GPS hooked up to a FC or OSD and had better precision.

The manual and firmware can be found here: https://www.frsky-rc.com/gps/

Thanks. I've also tried contacting Aloft Hobbies and they aren't sure either. I've seen some videos describing this sensor and there hasn't been any discussion or explanation (that I can find) related to the seemingly low precision. I'll keep tinkering and report back.

 

[Richrfl]

Dylan, I am having a headache with the FrSky GPS, besides some discrepancies in Altitude and Speed readings, but the one that most annoy me is the lag of 5 seconds in the data logger. Did you notice the same problem or do you know how to fix it? FrSky support never answered my enquire about that and the "manual" indicates "real-time", which in my book is way less than 1 second.

 

[dylanbeaudette]

Hi,

It seems like there are other reports out there on the internet of similar problems (lag), but I've not heard anything else about the limited precision and discrepancies (speed / altitude). As with anything based on GPS, there will always be some "flutter" in the readings.

So far, it seems like the approximate speed and altitude values are sensible... but I have no other way to measure directly, mostly estimation and sensing relative differences. I suppose that one could do some experiments in a field, after converting the raw telemetry into linear distance.

This video has some of the best suggestions / discussion that I've been able to find:

I tried increasing the logging rate to 0.1 seconds, but still see a loss of precision in the logged GPS coordinates. This series of flights was over a full size track / football field.