EKF with Rover 2.50

Posted by David Boulanger on February 3, 2016 at 1:56pm

Anybody using the EKF with Rover 2.50.  I had a couple of meters GPS glitch for about 30 seconds today and was wondering if the EKF was enabled would I have held a better course.  It caused no problems I'm just wondering if I should be using it or not.  I was repeating a waypoint mission 5 times in fairly close quarters around buoys.  All was good except for a brief period of time. 

Regards,

David R. Boulanger

You need to be a member of diydrones to add comments!

Join diydrones

Email me when people reply –
Follow

Replies

  • Admin
    Thomas J Coyle III February 3, 2016 at 4:04pm

    Hi David,

    I have used EKF1 very successfully with Rover firmware v2.50 and dual GPSs.

    EKF is far better than DCM in my subjective opinion based on actual testing on the same course using both an APM2.6 and a Pixhawk 1.

    If you are getting GPS glitches I would recommend a second GPS that is as far from you RFD900 as possible or reduce your RFD900 transmitter output slightly.

    Regards,

    Tom C AVD

    • David Boulanger > Thomas J Coyle III February 3, 2016 at 4:32pm

      Tom,

      Thanks for the reply.  It was my first test with my new Pixhawk and GPS.  I was on the water for 1 hour and had 30 seconds of a problem.  My RFD900 is set at 1/2 Watt (27 db. ) and about 3 1/2 inches from the GPS just like it was with my APM 2.5.  I have not looked at my log yet but will in the near future.  First glitch I have ever noticed.  The water today was very choppy causing the boat to roll and pitch a lot.  I ran the same course about 15 times trying some different parameters and that was the only glitch I noticed. 

      Regards,

      David R. Boulanger

    • Admin
      Thomas J Coyle III > David Boulanger February 3, 2016 at 4:37pm

      Hi David,

      Pitching and Rolling can cause the GPS to loose lock and if it was a somewhat cloudy day, the satellite signals can be weakened.

      As you have said, it is best to look at the GPS satellite count and HDOP to see where the loss occurred.

      Regards,

      Tom C AVD

    • David Boulanger > Thomas J Coyle III February 4, 2016 at 8:03am

      I looked at my logs today.  Sat count was 12 - 14 and a HDOP of 1.3 for the most part.  The bread crumbs show the boat going a meter or two to the left and the boat actually went to the right which I would expect.  I'm not going to worry about this right now until I have calmer water.  I did notice something interesting with the compass variance however.  When in AUTO my compass variance was for the most part good, but when I was not moving through the water very fast or close to shore I was getting some errors and spikes on the variance graph.  Here's what I think was happening.  The lake was very choppy.  The waves were causing very sharp movements to the boat in roll, pitch, up and down, yaw, and even laterally. This might have been confusing the sensors a bit.  Moving through the water in AUTO at 2.5 Knots the variance graph smooth's out and has no spikes.  Even with the variance error showing up on the HUD the boat tracked normal.

      One more thing.  I was looking at the EKF Parameters.  EKF GPS TYPE is currently 0.  I'm thinking with a boat this should be 1 because I operate in 2 dimensions only.  I'm I right?.

      For the most part, the conversion from my APM 2.5 to a PIXHAWK with the new Compass/ Mag module has gone well.

      Regards,

      David R. Boulanger 

    • Developer
      Grant Morphett > David Boulanger February 7, 2016 at 5:45pm

      Regarding the EKF_GPS_TYPE it shouldn't really matter as none of the navigation takes into account altitude.

      Thanks, Grant.

    • David Boulanger > Grant Morphett February 7, 2016 at 6:24pm

      O.K.

      Thanks Grant!

This reply was deleted.