20210929 Dev Meeting

Agenda

General:

• Update from relevant previous calls
• Tagged: issues, PR.

Notes

• common: allow any compid to be rebooted #1706 - better design, removes WIP.

  • Agree merge. Done.

• MAV_FRAME - better match descriptions to RFC #1700

  • Use of "BODY" for vehicle-fixed origin in local frame is clearly a big problem. Can we perhaps declare duplicate enum values with the same meanings and a better conventions. James suggested AHRS or IMU might be options (better than COG).
  • Dev call assessment - merge as it is (DONE) as this improves clarity significantly. Open issue/PR to further discuss the use of "BODY" as origin.

• common: add gimbal device joing angles extension #1704

  • Concern that this might be a "tip of the iceberg" indicating other missing information or flaws in the design.
  • Suggest separate message to allow easy iteration/prototyping.
  • AO on Julian - DONE: https://github.com/mavlink/mavlink/pull/1708

• Mission item grouping commands and messages #1687

  • Are group ids are globally unique or mission-unique, how will they be used, why do they need to be so big, how avoid clashes?
    • Matej - envisaging globally unique large ids mapped to locations via database that could be accessed by multiple GCS. We use a big ID to reduce the chance of clashes, but accept them if they happen.
    • Lorenz - A group indicates a feature of the mission. How you get to that feature matters - the same feature in another mission is a different group. In other words, group ids are mission specific. Also not happy to live with clashes.
  • Result:
    • Group ids are mission specific.
    • They should be allocated sequentially but do not have to be.
    • They are the size limited based on the max size of parameter in associated setter command - which is much bigger than it needs to be since most missions should have at most "tens" of features.
    • Mapping of group ids to features via mission can be achieved using MISSION_CHECKSUM.
      • Lorenz - MISSION_CHECKSUM design perhaps not optimal if it forces flight stack to generate the checksum. [is that avoidable if you have partial updates?]. But the approach can live with a mission item setting mission checksum so we can go ahead with it.
    • The group start/end message can include the associated mission checksum too.
  • AO Matej to add checksum and any docs to clarify usage. Will need to implement checksum.

• Add additional common camera parameters #395

  • Matej - has agreed to review

Attendees: JamesP, Seb, HamishW, JulianO, Matej, Lorenz