Dev Call April 19, 2021

Issues & Pull Requests




Google Summer of Code

1 Like

Attendee count (max): 19

UTC2301 -

  • GPS Galileo auto-configuration
  • Want to see by inspection that it doesn’t affect anything except the M8
  • Still seems to modify behaviour
  • Bumped to DevCallEU

UTC2302 -

  • Suppress speed scaling without AS sensor on takeoff
  • Only does its thing in mision takeoff and takeoff
    • Not FBWA
  • Merged!

UTC2307 -

  • KDECAN channel mapping attempt 2
  • New driver type so the new mappings won’t affect older configs
  • Old driver type will no longer work
  • Wiring is truly weird on these units - pay close attention when connecting these
  • Message needs attention - users must fix mapping before changing the types

UTC2312 -

  • Get water temperature from NMEA
  • Still works for normal data input
  • NMEA captures for all devices would be good
  • Randy will capture a NMEA file for this GPS
  • Tools/captures ?

UTC2321 -

  • Paul’s looked at it and called issue
  • Parameter conversion bug?
  • Tridge will look at this

UTC2325 -

  • Fly QNH alt in Plane
  • Baro gives qnh alt
  • Smooths things over with the EKF as qnh gets updated
  • When flying relative altitude mission points then they should switch between qfe and qnh
  • Glide-slope?
    • One waypoint inside radius, one outside
    • Is this really a user-interface thing that should be handled outside the EKF?
    • Interpretation of data
  • Baro library should only return function whether you’re using qfe or qnh and the delta
  • Treatment of user inputs rather than EKF?
  • Something flying fbwb crossing the radius may dive - by design
  • We already have baro-alt-offset crap
  • Paul tridge, Sam Andy and PH will chase this further

UtC2353 -

  • Copter: attitude target thrust-as-thrust bit moved to GUID_OPTIONS
  • Needs Wiki stuff
  • Just moves the option bit into a more relevant place

UTC2357 -

  • Quaternion attitude controller
  • Leonard’s been working with people to correct terminology
  • Control fixes - in particular fixing yaw errors on suboptimal paths
  • Thrust vector can be uniquely defined with heading with this PR
    • Allows for optimisation
  • Position controller has many years of patches and stuff in it which needs fixing
  • Heading is now the heading you’d have after your rotated level using roll and pitch
  • Currently if you get a big disturbance the euler angles will try to keep the yaw attitude with the same emphasis as roll/pitch
    • You try to maintain heading while correcting roll/pitch
    • Punching the rudder pedals
    • Disturbing the airframe
    • Over-controlling the aircraft
  • Paths between attitudes
  • Peter Hall: My understanding is that the shortest distance between the roll pitch and yaw angles is not the shortest distance between the rotations the endpoints represent.
  • [10:13 AM] Peter Hall: So this fixes that.

UTC0014 -

  • SCurve input shapers for alt hold

UTC0016 -

  • Add extension fields to GPS2_RAW
  • Can be merged on CI pass

UTC0022 -

  • Waf fixes needed for argument-list-too-long cygwin issues
  • GNU linker trick to fix problem?
  • Need to break it up into chunks
  • Master still compiling for Randy - but not for long?
  • ESC notch PR fails because of this

UTC0031 -

  • Make passthrough more reliable
  • Merged!

UTC0035 -

  • Reset_mode_switch?
    • Will be removed
  • Can be merged once that bit is removed

UTC0038 -

  • Merged
  • Increase stack thread on SmartAudio

UTC0039 -

  • AP rather than APM in mavproxy

UTC0041 -

  • Merged
  • Fixes mro compass ob Cube

UTC0043 -

  • Need to account for pitch/roll rate when determining attitude at GPS time horizon
  • Already have some circular reasoning
    • Uses gyo data rather than attitude solution to reduce the problem
  • 0.1 comes from measurements on F9P

UTC0055 -

  • Wants review!
  • Upload of lots of stuff via mavftp

UTC0056 -

  • Various reports of things not working
  • 3D backaway loiter for 4.1 would be nice
  • Sergey reports issues on iFlight Beast8
  • Seddd internal errors stack overflow using SmartAudio
    • Already fixed
  • 38km limit still needs to be looked at
    • EK3 issue
  • Thanks to the devs heling out with 4.1! Much appreciate

UTC0103 -

  • Several pending things that still need to be done for 4.1
  • AC_PID stuff is in
    • Autotune is much better
  • Quadplane landing approach code still to go in

UTC0108 -

  • Nothing to report for Rover

UTC0109 - GSoC update

  • Application period ended
  • PH has been looking at some of them
  • Can’t discuss on open forum
  • Fewer applications but that’s consistent across all organisations

UTC0112 - close

1 Like