Welcome to PhantomPilots.com

Sign up for a weekly email of the latest drone news & information

Satellite internet: possible GS hazard!

Discussion in 'DJI Ground Station' started by GearLoose, May 21, 2014.

  1. GearLoose

    Joined:
    Apr 4, 2013
    Messages:
    475
    Likes Received:
    1
    Tried again today to get a safe waypoint flight and made some progress, though it revealed that the root of the "off track" problem is probably our satellite internet service. For several months I’ve noticed a consistent error when viewing Google Earth, an “offset” of my tracks and waypoints of 160 – 200 feet. No help from Google and our internet provider didn’t have a clue, though they did suggest it might be a latency problem. That is the most likely cause but why did it start suddenly? They have no idea.

    Yesterday I almost crashed into a tree line when the Phantom took off in what appeared to be the wrong direction (GS mission). Today, just to be very safe, I set a 100 foot on-a-side triangular waypoint mission. As before the Phantom took off properly, then went right when I expected it to go left.

    I then set a single waypoint about 100 feet away and 80 degrees to my left. The Phantom flew away to the waypoint but at about a 20 degree angle. That was enough to convince me that the Google Earth data I'd cached using our satellite internet has to be seriously affecting these missions.

    I was flying again at a neighbor’s pasture and it turns out he’s having a similar problem with Google Earth — but his provider is Dish, whereas ours in Wildblue — completely different satellites.

    So, tomorrow I will find “real” internet and cache the Google Earth maps on my iPad again. Not exactly convenient but better, I guess, than trying to calculate the error factor in order to adjust the mission tracks.

    I'll post again once I've made more tests but in the meantime, be very careful if you're using satellite internet to prepare your Ground Station missions.
     
  2. sar104

    Joined:
    Mar 22, 2014
    Messages:
    1,265
    Likes Received:
    245
    Location:
    Los Alamos, NM.
    I read that several times and I have to admit I have almost no idea what you are hypothesizing.

    Are you suggesting that the coordinates of the waypoints that you are setting are incorrect? If so, how could that be a problem with your internet provider? Google's maps are all fully geo-registered to within a few feet at worst, independent of how you get them onto your computer or iPad. I cannot imagine what IP latency has to do with it.

    If your recorded tracks are displaying incorrect on Google Earth there are several possibilities. Your GPS track data could be inaccurate for some reason (local interference, poor reception, damaged antennas or receiver etc.). You could inadvertently not have locked the track to ground level in Google Earth, and so are ending up with a parallax error on viewing. Or Google Earth could be a couple of hundred feet out, which is very unlikely almost anywhere on this planet.

    Or am I completely misunderstanding your problem?
     
  3. GearLoose

    Joined:
    Apr 4, 2013
    Messages:
    475
    Likes Received:
    1
    Let me put it this way: if you record a track with a Garmin GPS you know to be accurate, say from your home to a prominent landmark... and you upload that track to Google Earth and find that it shows your track to be anywhere from 50 to 300 feet "off"... that would illustrate my problem. Here's a screenshot from GE of a waypoint I marked on a bridge a couple of years ago.

    Believe it or not, that waypoint actually appeared on the bridge in GE... two years ago. Now it is over 100 feet off.



    A second screenshot shows a local road. Google Earth's satellite image shows the road clearly but GE itself shows the road as anywhere from 50 to 300 feet from its actual track.



    I've been using a GPS since they first came out for hikers. I've never had a problem like this before -- and this problem only showed up in the last few months, both for me and for my neighbor. I live deep in the sticks so I don't have anyone else to compare notes with but something is clearly wrong here. The only common link between the two of us is satellite internet -- he doesn't use a GPS.

    So what I'm suggesting is that Google Earth and/or my satellite internet is the culprit in the unexpected deviation I saw over the past two days when I tried to fly waypoint missions from my iPad.

    What do you think?



     

    Attached Files:

  4. zenith

    Joined:
    Apr 22, 2014
    Messages:
    35
    Likes Received:
    0
    is your phantom's position correctly represented in the google earth maps?

    google earth might have an offset but it usually happens only in china.
     
  5. GearLoose

    Joined:
    Apr 4, 2013
    Messages:
    475
    Likes Received:
    1
    I'm not really sure -- the place where I fly is very uniform and the GE image isn't detailed enough to distinguish exact spots. But, I'm probably within 50 -100 feet of the indication position of the Phantom.

     
  6. sar104

    Joined:
    Mar 22, 2014
    Messages:
    1,265
    Likes Received:
    245
    Location:
    Los Alamos, NM.
    OK - I see the problem. A few followup observations and questions:

    (1) It cannot have anything to do with your internet provider, which has absolutely no effect on the actual data that you download, mapping or otherwise.

    (2) How, exactly, does your neighbor see this problem without a GPS?

    (3) How do you know that your GPS track is accurate on a road with substantial tree cover?

    (4) Which coordinate system and datum is your GPS reporting in? Probably lat/long rather than UTM, since Google Earth only very recently began supporting UTM, but is it using NAD27 or NAD83/WGS84? There is a significant difference between those in many areas, and that north/south offset could possibly be a datum shift.

    (5) It seems unlikely that GE would have those kinds of errors, but it is simple to compare with a known reference map such as USGS. I would have to be able to identify that location to be able to check.
     
  7. GearLoose

    Joined:
    Apr 4, 2013
    Messages:
    475
    Likes Received:
    1
    the neighbor sees the same errors shown in my screenshot from GE of the local road -- there is a parallel track to the actual road.

    I use a Garmin 60Csx In the past it has been remarkably accurate -- I have many tracks recorded on open logging roads, with no tree cover. Previously, GE would display these tracks right on top of the roads shown in the satellite image. Here's an example of bicycle tracks that were previously very accurately displayed in GE.



    Furthermore, the same problem appears in tracks I record using the FlyTrex gps logger. Last year, for example, my FlyTrex tracks (uploaded to their site, then exported to GE) were obviously quite accurate. Lately, however, they are also off -- and by the same margin, usually 60-200 feet.

    I don't see how the coordinate system could be an issue on tracks that were recorded in past years but now display incorrectly on GE?

     

    Attached Files:

  8. sar104

    Joined:
    Mar 22, 2014
    Messages:
    1,265
    Likes Received:
    245
    Location:
    Los Alamos, NM.
    I agree that the 60CSX is a fairly accurate unit. The only obvious way to proceed is to check the GE coordinates (such as the middle of the bridge) against a known standard, but without knowing the South Fork river in question I can't do that.

    The tracks that were previously correct in GE - are those saved locations in GE that were previously looked correct and are now wrong, or did they look wrong when you reloaded them to GE recently from the GPS unit?
     
  9. GearLoose

    Joined:
    Apr 4, 2013
    Messages:
    475
    Likes Received:
    1
    Yes, the tracks that now display incorrectly were uploaded to GE in past years. Until recently they displayed correctly. I haven't tried re-loading those to GE.

     
  10. GearLoose

    Joined:
    Apr 4, 2013
    Messages:
    475
    Likes Received:
    1
    I've just completed another test: using my Garmin 60CSx I recorded two local waypoints at places that are clearly visible in Google Earth satellite images.

    Having disconnected from the internet, I uploaded the waypoints to Garmin Basemap (mapping software). Basecamp displayed the waypoints exactly where I'd recorded them -- easily within a meter of the exact spot.

    I then switched Basecamp from map view to Google Earth view. Both waypoints were just over 300 feet to the south of their actual recorded location.

    So "SAR104", doesn't this indicate that my Google Earth/Ground Station problem is caused by my internet connection?
     
  11. sar104

    Joined:
    Mar 22, 2014
    Messages:
    1,265
    Likes Received:
    245
    Location:
    Los Alamos, NM.
    No. That quite clearly suggests that Google Earth's georegistration is incorrect in your area, either simple error or datum mismatch. As I said before, the only way to know for sure is by coordinate and datum delta comparison at your location, but without knowing where you are I cannot distinguish between those possibilities.

    I'm still baffled how you think that an internet connection can change the downloaded map data.
     
  12. GearLoose

    Joined:
    Apr 4, 2013
    Messages:
    475
    Likes Received:
    1
    I suspect the internet connection because I don't know anything about "georegistration" and am desperately casting about for a solution!

    Since my last post I've compared the actual lat/long figures, so here they are:

    Garmin:

    N48.75484
    W122.12713

    The same waypoint exported directly into GE:

    48.754844°
    -122.127132°

    Measuring from the waypoint icon to the actual spot where I recorded the waypoint, the distance is 300.89 feet and the heading is 1.68 degrees.

    Is there anything else I can provide that might be helpful?

    thanks!



     
  13. sar104

    Joined:
    Mar 22, 2014
    Messages:
    1,265
    Likes Received:
    245
    Location:
    Los Alamos, NM.
    OK - that I can work with. It turns out that Google Earth has a coordinate offset in your area of about 360 ft at 3.8 degrees by comparison with reference points on the local USGS quads (Deming and Canyon Lake 1972 revisions), and about 275 ft with the latest USGS 2014 Canyon Lake quad and with Bing maps. Interestingly the discrepancy vanishes to nearly zero on moving 8 miles east and comparing to the 2012 USFS quad (Cavanaugh Creek), which at that point then has a 75 ft discrepancy between the 2012 USFS and the 1989 Cavanaugh Creek USGS quad.

    It appears to me that there is substantial variation in georegistration accuracy in your area, and it is possible that Google made adjustments in the last couple of years that resulted in a substantial error on those quads.

    Not sure what to make of that, other than it looks like some remarkably sloppy mapping. Not your internet connection, anyway. I would avoid GE for mapping and use Bing maps instead, but keep in mind that waypoints that you generate from such maps will be accurate to around 50 ft at best.
     
  14. GearLoose

    Joined:
    Apr 4, 2013
    Messages:
    475
    Likes Received:
    1
    In at least one sense I'm quite relieved (and grateful) that you've found the root of the problem -- but I fear that it will make using the Ground Station impractical in my immediate area. I had hoped to map the local wetlands but they meander a great deal and are surrounded by tall forest. Perhaps there is a way to map certain places with my Garmin, then transfer exact coordinates to GS waypoints?

    I've never used Bing maps so I wonder if they can be accessed from the DJI GS app for IOS, instead of Goggle Earth's?

    Again, I thank you for the help and explanation of the problem. I'll pass the information on to my neighbors -- and Wildblue also asked me to contact them if this eventually got sorted out.

     
  15. sar104

    Joined:
    Mar 22, 2014
    Messages:
    1,265
    Likes Received:
    245
    Location:
    Los Alamos, NM.
    It doesn't look like there is any way to change maps in the GS app. However - having looked at the app, it does not appear to be using Google Maps, it's actually using Bing maps and it's coordinates agree with the local topos. If you ask it to display your location on the aerial map layer, does it get it right or do you see the same 300ft error? If it is correct then you can just add your waypoints directly on the map, and those should be correct in flight.

    Similarly, if you add a waypoint on the app map at your position, do the displayed coordinates agree with your GPS (check that the GPS is in WGS84/NAD83 datum though)?
     
  16. GearLoose

    Joined:
    Apr 4, 2013
    Messages:
    475
    Likes Received:
    1
    As you suggest, sar104, I've done a test by comparing waypoints set in the iPad GS app with waypoints at the same locations set with my Garmin, with Motion GPX app, and with Google Earth.

    The only waypoints that show any significant variation are set by Motion GPX (which must be using Bing or some other satellite imagery as the views are closer and sharper than any others).

    For example, here are the latitudes:

    Motion GPX 48.4507
    Garmin 48.75109
    Google E 48.751091
    iPad GS 48.7510971753

    I hope I'm correct in interpreting this as confirming the accuracy of the iPad GS app? If so, I still have to figure out why my actual test flights didn't go as expected -- but my prime suspect there would be some form of pilot error. :)


     
  17. sar104

    Joined:
    Mar 22, 2014
    Messages:
    1,265
    Likes Received:
    245
    Location:
    Los Alamos, NM.
    Just to clarify - how are you setting those waypoints? Are you placing a marker on an aerial image displayed in each of those apps and then looking at the lat/long of the resulting waypoint?

    And I think that you may find that Motion X is simply using degrees, decimal minutes, rather than decimal degrees. 48˚45.07' is the same as 48.75109˚.
     
  18. GearLoose

    Joined:
    Apr 4, 2013
    Messages:
    475
    Likes Received:
    1
    Yes, other than the Garmin's waypoint, which I transferred to Basecamp (and then used "view current map in GE", I manually placed waypoints and then copied their coordinates.

    Once it is displayed in GE (via BaseCamp) the Garmin's waypoint icon is 300.89 feet off the mark, almost due south. But, the coordinates shown in the info window are those I copied above.

    I just went back to GE and placed a waypoint on the satellite image to match the actual location, removing the 300' error. The coordinates of that waypoint are:

    48.751849°
    -122.112696°

    So Motion X is actually in agreement with the others -- I never would have puzzled that one out.

    Can you draw any conclusions from this yet?



     
  19. sar104

    Joined:
    Mar 22, 2014
    Messages:
    1,265
    Likes Received:
    245
    Location:
    Los Alamos, NM.
    Yes - as I mentioned previously, you are seeing the same discrepancy between Google maps and others that I saw when I checked - of the order of 300 ft south. The only thing that I cannot do remotely is take actual GPS coordinate readings in the field and check which map sets they agree with, but my suspicion is that Google maps has the largest error.

    My suggestion is still that provided you find that the Groundstation app (which uses Bing aerial imagery) displays your location correctly (if you physically go to a location and then look at the app), you should be able to set waypoints in the app that are reasonably accurate. If you don't have internet and can't actually see the aerial imagery when you are at the location, just set a waypoint for your precise location and then check when you get back to see if the waypoint appears to be in the correct place. You can ignore that the locations and tracks look wrong in Google Earth or maps.
     
  20. GearLoose

    Joined:
    Apr 4, 2013
    Messages:
    475
    Likes Received:
    1
    OK... understood. Once it stops raining I have to fence my blueberries today but with any luck I'll be able to do a waypoint flight test tomorrow morning.

    thanks again for your patience and help.