Satellite internet: possible GS hazard!

[GearLoose]

The rain stopped so I set up a test in the driveway, marking a single waypoint in the iPad GS app at the place where I was sitting with the Phantom. This is the same place I'd marked earlier with a waypoint in Google Earth, my Garmin, and Motion X.

Driveway: iPad GS, sitting at the waypoint:

122.11274
48.75109

Driveway: iPad GS waypoint placed earlier, at some distance from the actual spot.

122.11273
48.75109

I did a similar comparison with a couple of other waypoints around the property -- again, the difference in coordinates seems very small.

One more wrinkle: the GS app showed the aircraft as 117 feet away when it was actually sitting on the waypoint. When I took my Futaba for a short walk, testing a new fpv antenna, the iOSD Mini also showed me as being 34 meters from the Phantom -- again, when I was actually next to it.

 

[sar104]

That's good, but the significant question is - when you were in your driveway, did the iPad GS app show you in your driveway on the aerial imagery? Or did you appear to be off in the bushes somewhere several hundred feet away?

 

[GearLoose]

Yes, the iPad GS was "spot on", showing the waypoint right where I was sitting -- but the aircraft's icon was out over the wetlands, over 100 feet away.

That's good, but the significant question is - when you were in your driveway, did the iPad GS app show you in your driveway on the aerial imagery? Or did you appear to be off in the bushes somewhere several hundred feet away?

 

[sar104]

OK - so the iPad's internal GPS was showing you in the correct location on the app (Bing maps), but the aircraft was reporting a location 100 ft away. How many satellites was it receiving at that time?

 

[GearLoose]

Unfortunately, I didn't note the number of satellites.

OK - so the iPad's internal GPS was showing you in the correct location on the app (Bing maps), but the aircraft was reporting a location 100 ft away. How many satellites was it receiving at that time?

 

[davipt]

Google's maps are all fully geo-registered to within a few feet at worst...

That's not quite true.
As a major contributor for the Waze product, I've been hearing horror stories and suffering myself with this subject. We do get GPS tracks from a lot of drivers, hence easily discarding bad GPS devices or GPS signals affected by mountains or other stuff.
What we see is that the quality of the map varies a lot with the country and even with the area of the country. The majority considers BING maps (or the aerial maps used by everybody else) with much more quality, more recent, and quite aligned. The complains from the street map editors usually were only on cases where the maps had not enough quality.
Since Google bought Waze and the aerials changed, the stories have been mixed. In some countries the maps became better, but mostly because the bing one had no quality.

On the countries where bing was perfect, google is now a piece of crap. And the problem is that the aerials aren't even consistently misaligned. They may be 20m off east in one place, and 50m off north-west just some kilometers away.

I've seen myself cases of errors of more than 100m. One I've tested myself as there is a GeoCache nearby.

Long story just to say "don't trust google maps". Double check with other providers. Try to get some quality local maps if available. And at the end, give it a huge margin from whatever you're seeing on the aerials. And don't forget the aerials are always old, maybe there is a new building or cable or tree on the real world today!

 

[sar104]

Google's maps are all fully geo-registered to within a few feet at worst...

That's not quite true.
As a major contributor for the Waze product, I've been hearing horror stories and suffering myself with this subject. We do get GPS tracks from a lot of drivers, hence easily discarding bad GPS devices or GPS signals affected by mountains or other stuff.
What we see is that the quality of the map varies a lot with the country and even with the area of the country. The majority considers BING maps (or the aerial maps used by everybody else) with much more quality, more recent, and quite aligned. The complains from the street map editors usually were only on cases where the maps had not enough quality.
Since Google bought Waze and the aerials changed, the stories have been mixed. In some countries the maps became better, but mostly because the bing one had no quality.

On the countries where bing was perfect, google is now a piece of crap. And the problem is that the aerials aren't even consistently misaligned. They may be 20m off east in one place, and 50m off north-west just some kilometers away.

I've seen myself cases of errors of more than 100m. One I've tested myself as there is a GeoCache nearby.

Long story just to say "don't trust google maps". Double check with other providers. Try to get some quality local maps if available. And at the end, give it a huge margin from whatever you're seeing on the aerials. And don't forget the aerials are always old, maybe there is a new building or cable or tree on the real world today!

Thanks for the information - I was previously unaware of this issue until I started looking into this particular case. What you describe is consistent with what I found for this locale.

 

[GearLoose]

I finally completed a successful waypoint flight today!

Taking a cue from sar104's recent posts, I invited a friend with a strong background in GPS, radio telemetry and mapping to join me. (He also flies a Phantom.) I first set a triangular mission with legs about 100 feet long, just as I'd done a few days ago with such disappointing results (Phantom meandering badly, failing to track).

In today's test the Phantom did make each waypoint but it did so almost drunkenly, with what looked very much like toilet-bowling en route. Then, reaching WP 1 on the second lap, it veered away to the west, clearly out of GS control. (I regained control in ATTI, once again forgetting to increase the throttle, which resulted in a crunch-and-go landing and another broken landing strut. Live and learn... eventually. )

At this point my friend got out his GPS and noted that the narrow valley we were in didn't have very good satellite coverage. He speculated that poor GPS coverage combined with the reduced area between waypoints (and a gusty wind) was basically confusing the Ground Station.

To test this idea I doubled each leg of the triangular mission to 200 feet.

Sure enough, the Phantom flew the mission and appeared to behave itself quite well.

 

[ianwood]

Just skimmed this, geo-registration errors are common but most often they're pretty small. Atmospheric drift is a pretty likely factor in GPS track errors. Your satellite provider is not a cause of interference. If it is, call the FCC as they have shut down entire businesses over GPS interference.

 

[sar104]

Just skimmed this, geo-registration errors are common but most often they're pretty small. Atmospheric drift is a pretty likely factor in GPS track errors. Your satellite provider is not a cause of interference. If it is, call the FCC as they have shut down entire businesses over GPS interference.

In this particular area it turns out that there are significant discrepancies between Google, Bing and USGS. I was very surprised.