Moving Base and EKF Variance Issues

ArduRover Rover 4.6
1

Greetings (I’m a newbie and not sure if my photo and param files attached),

Issue

  1. When I try to ARM the rover I get EKF variance failsafe messages
  2. Current config seems a little flaky to get RTK Fixed. GPS2 will get RTK Float and GPS1 will get 3D dgps. Not sure why getting RTK Fixed is an issue. RTK Fixed was easy to obtain when it was a single GPS and when I send NTRIP to “all” under GPS_INJECT
  3. I think that I have moving base working because when I move the rover, the MP red line shows the proper orientation on the map.
  4. EKF Status shows pos_horiz_abs as “off” and red

My Unconfirmed Suspicion

  • The FC is getting mixed info from NTRIP and the moving base sending mixed location messages to the FC

Request/Questions

  1. Pls look at my attached config to if moving base looks correct
  2. Found mixed info online about setting GPS_DRV_OPTIONS to 1. Should I?
  3. Does this seem to point to incorrect RTCM messages incomplete or to the wrong GPS?
  4. Any other items to review for EKF variance?

My Setup

  • ardupilot rover skidsteer version 4.5.7
  • brushed wheelchair motors 24 volts
  • cube orange plus
  • cytron smartdriveduo-30 mdds30 (independent mode)
  • rc input pins from cube orange servo 1 and servo 3
  • RC1 pin connected to servo 1
  • RC2 pin connected to servo 3
  • moving base ublox where gps1 is the base and gps2 is the rover
  • external ntrip feed to the rover
  • gps1 has a single wire going from the SimpleRTK2B Rx UART2 to the SimpleRTK2B at gps2 Tx UART2
  • FC is in the middle/top black box and the motor controller in the bottom one
    image
    image768×900 96.8 KB

    3-12-25.param|attachment (16.4 KB)
1 Like
2

Hi @ronzika,

So it sounds like the EKF doesn’t have a position estimate

EKF Status shows pos_horiz_abs as “off” and red

It’s hard to know the cause without a log file and maybe the param file was not attached correctly?

3

This is possibly part of the problem (but maybe not the whole story). Your words are correct based on last known weird silk screening of those boards, but it’s very easy to make an error here.

Disable the GPS_DRV_OPTIONS bit and see if things improve.

May help to troubleshoot one thing at a time. Moving baseline (gps for yaw) should work without NTRIP.

As Randy says, reattach the parameter file and/or log file.

4

3-12-25.param (16.4 KB)

Thanks guys! File attached this time (I think).

What was your thought on that being part of the problem? Thinking it may be reversed? Am I correct that you can either do the serial wire or back through the FC?

Thanks,
RZ