Servers by jDrones

Dev Call Nov 2, 2020

Issues & Pull Requests




1 Like

Attendee count (max): 19

UTC1100 -

  • Fix usage of rangefinder in landing
  • Bad landing slope with invalid data
    • Even if rangefinder is not enabled
  • Two places where bad rangefinder data could lead to bad landing
  • Will go into next Plane 4.0 stable
    • Next Plane will also include fixes for Pixracer Pro pins

UTC1108 -

  • Sub-classes for heli autotune
  • And add a sequence thing
  • Testing?
    • Autotune starting with default parameters on Calisto model in master
    • Does it get similar tuning results?
    • Already a bunch done
  • No heli autotune in here, just some reorganising for same….

UTC1120 -

  • Removes a race condition
  • Some sort of time check?
  • Preference for attitude.cpp not to be polluted by tailsitter stuff

UTC1137 -

  • Tailsitter gain scaling parameter goes away
  • Azure gone crazy again

UTC1138 -

  • Merged

UTC1139 -

  • Can be merged once Peter and MdB are happy

UTC1142 -

  • Output dspoilers etc even if not fully configured
  • Avoid user confusion
  • Prearm check that all are configured?

UTC1144 -

  • Add a Chinese font
  • Why are we carrying 4 English fonts @3kB+ each?
  • We handle fonts going away gracefully
  • We got four fonts from early on
  • Should be easy to pull from SD card?
  • Use apjtool to replace a font?

UTC1157 -

  • Approved by Paul
  • EKF stuff
  • Merged

UTC0000 -

  • Helpful to find constrain_nan bugs
  • Only enabled on 2MB or larger flash
    • 1.5kB cost
  • Merge this now and perhaps make it across all boards later

UTC0004 -

  • Make parameter show up correctly
  • merged

UTC0006 -

  • Third time around
  • Minor fixes required then will be merged
  • Better handling of a primary gps sensor

UTC0007 -

  • Merged!
  • Tested on Septentrino hardware and in SITL
  • RTK baseline stuff

UTC0012 -

  • Randy to look at

UTC0014 -

  • Arming in stabilize without position yields ekf failsafe ATM
    • After this you will get tones as well…
  • Shouldn’t allow EKF failsafe if we’ve never had a position estimate
  • Forced arm in loiter mode?
    • Not permitted
  • Will hold off on this until the failsafe issue is fixes to avoid annoying users…

UTC0018 -

  • Don’t ask GPS, ask the AHRS
  • We’d need to know each of these checks is going to do the right thing
  • If you don’t have GPS lock then this will give different results
  • We don’t want it to work without GPS…
  • Without accel cals bad ground velocities can be had
  • Fault
    • If GPS gives bad velocity its ublox’s fault, if we use AHRS its more our fault…
  • Generally Randy things we should be using AHRS more (like this PR)
  • Randy doesn’t see cases where the EKF says it has a good velocity estimate and gives you wildly bad values
  • Why?
    • GPS-denied testing
    • Not dead-reckoning when we could be
  • Tom will go and have a look at his log files and work out how he managed to auto-takeoff without a GPS (apparently)
  • Airspeed estimates are often garbage on the ground

UTC0038 -

  • References on non-boolean returns
  • Scripted LEDs!
  • Merged!

UTC0040 -

  • Paul is reviewing now it is a manageable size
  • Additional check coming in for where all sources are None
  • Before or after Replay?
    • Probably after as then we can test this PR
  • This will need to be changed after Replay goes in
  • Randy would love to help people test this
    • Two sources
    • E.g. Wheel encoders + GPS
      • Drive Rover through tunnel
        • Use bendy-ruler while doing this
    • Could use this on Planes…
  • GPS + optical flow + a script to use when lidar in range
  • Optical flow for dead reckoning in Plane with some patches
  • Terrain database + baro alt to get height-above-ground then use flow for velocity
  • Long-distance dead-reckoning….
    • Visual stuff

UTC0049 -

  • Can be merged once Mark and Peter are happy

UTC0053 -

  • Can be merged once the commit message is fixed
  • Might be a better solution for all of this past Randy’s sources PR

UTC0058 -

  • Nate’s issue was the only one remaining
  • Release this-arvo or tomorrow

UTC0059 -

  • Needs some doc fixes
  • Parameter inbound for GCS timeout
  • People no longer able to continue mission in auto mode?!
    • FS_OPTIONS allows it

UTC0104 -

  • Logging retrieval issue
  • If you get masses of logfiles you may then be unable to download logfiles
  • A partner managed to get a desired lean angle in poshold which survived a disarm/arm cycle

UTC0108 - KDECAN issues?

  • Randy can look at it with Tom’s hardware

UTC0108 - user alert issue?

UTC0112 - close

Servers by jDrones