Dev Call Apr 6, 2020

**Issues & Pull Requests"




1 Like

Attendee count (max): 23

UTC1102 -

  • More documentation
  • Randy will merge

UTC1103 -

  • Bug for scripting
  • Wrong number of arguments for some scripting bindings
  • Commented out code is a little ugly, MdB will fix in master
  • Set-mode binding is unsafe
    • Remove it?
  • More scripting stuff back into stable?
    • Lots of improvements in AP_Vehicle required
  • merged!

UTC1107 -

  • Copter add logger documentation
  • Time stuff needs fixing
    • Separate PR
  • Merged!

UTC1110 -

  • Valgrind?
    • Not yet
  • Mavlink message would be good to get in
    • merged
  • Loads of enums
  • Still need to stare at this to get correctness
  • And run it under Valgrind

UTC1116 -

  • Compass initialised vs enabled
  • Setting compass in the EKF
  • Need to test failing a compass at runtime
    • One compass fails
      • Should fail over to second compass
    • Two compasses fail
  • Use_for_yaw should check the _enabled
  • Ad RC switch to kill compasses
    • Driver keeps going API level forbids access to data

UTC1137 -

  • HAL_SITL ensure unique-id
  • Merged!

UTC1140 -

  • A few patches to come from Peter’s review
  • Improve support of T265 for non-GPS navigation
  • Thien is working on these
  • Sending mavlink messages into AP
  • Some special things in the camera we can use
    • Prearm checks to match orientation
  • Used to have point vehicle North at power-on!
  • Camera gives you earth-frame xyz
  • Fusing earth-frame yaw only - not rp
  • More tests for maths stuff coming
  • More documentation for running the examples under SITL would be good
    • ./waf configure --board sitl --debug
    • ./waf --target examples/rotations
    • build/sitl/examples/rotations -C --model quad
  • Fusing accel data in with no GPS data was a problem in the past for retaining good attitude estimate - will we have the same problem?
  • Improving the existing autotest would be good
    • Mucking up the alignment in the simulation would be nice
      • Setting angles in the built-in-SIM_Vicon.cpp
  • Would the camera work pointing straight down?
    • Yes
    • Any of them should work
    • forward/down/right all bench-tested OK
    • Don’t lean over by >45 at startup
  • VisualOdom now looks after the mavlink messages
  • … and now all vehicles have VisualOdom
  • Why have the old message?
    • Because some systems may not provide attitude
    • Calling it T265 isn’t a great name for the class
      • Can rename it later when we have another example
      • Not enough examples to work out what’s common and not common
      • Change behaviour if we see any change in the attitude message?
    • Mavlink2 extensions to provide confidences
      • Check with Rob Mahoney as to whether the covariances make sense
        • Or the guy from intel
  • It is behind minimise features
    • Lose support for vision-position-estimates you used to get
  • Ifndef’d?
  • Works on submarines now…
    • Velocity in three axes from Sonar

UTC0007 -

  • Can be merged after changes have suggestions made
  • Slow SITL down to accommodate slow mavlink packet

UTC0024 -

  • Avoid use of constants through code
  • Need to make sure code size doesn’t change
  • Comments required in code

UTC0031 -

  • Soaring upgrades
  • Marco’s been doing a lot of flying on this one
  • Rebased a lot
  • MdB hasn’t looked at the bit he’s supposed to
    • The line-up stuff with the loiter controller is critical
  • Tridge is otherwise inclined to merge it if it only affects soaring
  • Needs to be valground
    • Uses internal physics, so that’s good
  • Mission asks GPS for ground course which is odd, but pre-existing

UTC0040 -

  • Merged!
  • About 70 to go

UTC0042 -

  • Check that the sticks are neutral
  • Revisit from last week
  • Has_valid_input returns true if you have failsafes disabled
  • For arming check you now need to get an update in the last 500ms
    • Just add this to has_valid_input?
      • Would become a behaviour change
  • Mode1 vs mode2
    • Hard to arm with narrow deadzones?
  • Put it in and see how painful it is?
    • One extra options bit to turn this off
      • Can turn it off after a release is made that way

UTC0052 -

  • EKF2 skipping axes
  • Peter still to look at this

UTC0054 -

UTC0100 - Plane update

  • Nothing much from last week
  • Probably a new minor release coming
  • A couple of backport PR

UTC0102 - Copter update

  • Couple of new support issues
    • Reoccurence of parameter reset issue
      • Quite confident the bootloader fix was an actual fix
      • Randy has seen two resets in the last week
  • Lightware says SF30c doesn’t work with 4 does with 3
    • Randy’s getting a copy

UTC0102 - Rover update

  • No updates

UTC0104 -

  • Getting waypoint data in lua scripts
  • Next week

UTC0106 -

  • Speed estimate into EKF
  • Next week

UTC0107 - GSoC

  • Application period has closed
  • Potential mentors need to review
  • 2 or 3 weeks to look at applications

UTC0108 - close