FW/VTOL Call: December 6, 2022

December 6, 2022

Tracked issues from last meeting

Control allocation issues for fixed-wing

Failure detector - roll/pitch

  • @chris1seto couldnt reproduce, let’s call it done for now

Acro mode

  • @chris1seto tested - integral = bad.
  • @tstastny @sfuhrer @dagar check latest log to make sure there is not a bug in yaw rate integration
  • smooth out manual flight
  • or stabilized can do this…
  • most off the shelf RC planes have roll and pitch damping for acro
  • @sfuhrer either full rate control - or switch off yaw rate control
  • “safe mode” “acro mode” “rate mode”
  • what we actually want: tuning mode? or “fun” mode?

Quad-tailsitter SITL model flies poorly

Unifying rate controller

L1 purge

  • DO IT

Discussion

Control allocation

  • model feed-forward is not yet included in allocator - it only considers the “net torque” and wants to generate a torque
  • per channel scaling exists if we write the function in CA (e.g. how to map “spoilers” to existing control surfaces)
  • @chris1seto @ryanjAA CA may not be as easy as mixer

High fidelity sim for fixed-wing

Concerning yaw rate potential issue: I replied in the PR, didn’t find any issue in non-Acro mode: Fixedwing reset rate integral in acromode by Jaeyoung-Lim · Pull Request #20486 · PX4/PX4-Autopilot · GitHub