Skip to content

20211124 Dev Meeting

Hamish Willee edited this page Nov 25, 2021 · 11 revisions

Agenda

General:

Notes

  • Proposal: Support for common flight modes #1750

    • Proposal considered good baseline for discussion
    • Appears to have no side-affects - i.e. could co-exist with older systems.
    • Mode in HEARTBEAT as extension field preferred over separate mode message.
      • Various options like using range in custom mode field discussed. Probably not feasible, or prone to implementation risk.
      • Considered risky just because HEARTBEAT is so fundamental.
      • Mode as discussed is autopilot specific, but heartbeat is used by lots of things. Might be benefit in making heartbeat reflect modes for component too.
    • JamesP - get some broad feedback from ArduPilot
    • AO: HamishW create RFC to encourage better engagement.
  • new message radio_status_extensions in development #1748

    • We don't know enough about the subset of data suggested - what is provided by the API for their radio. What about other radios
    • What about multiple radios or MIMO radios that support multiple channels? Ids
    • Need more expert technical review
    • AO comment to ask more questions: HamishW - DONE
  • AP: common: added voltage and current multipliers to BATTERY_STATUS #233 and 16-cell support #1747

    • Needs a lot more thought - this is already a heavyweight message on systems with many batteries
    • Lorenz - let's propose ambitious redesign that can co-exist for a short period. We keep on having to extend it, so we should think about future proofing.
    • AO next meeting to discuss design options.
  • dev: add new message for component capability #1724

    • Agree this meets requirements.
    • DONE AO HamishW to merge.
  • Add MAV_CMD_SELECT_TRIGG_CAM command. #1749 - selecting current camera in a mission.

    • Concerns
      • using component ids for addressing - ids are for routing, not identity/type
      • settings outlast the mission.
      • No way for offline planning unless you know the vehicle.
      • Perhaps not the right abstraction - rather than specify the camera, perhaps should specify the function/stream and let the underlying system select the right camera.
      • We keep coming back and doing minor patches to this. Need to start with requirements of camera and gimbal together and make sure that we have the right architecture to deal with this.
    • Long and broad discussion about issues, abstractions, routing of vehicle specific behviour outside of MAVLink, getting proper broad input into design of the fundamentals.
    • General agreement we need to discuss this in more detail, but perhaps in dedicated meeting.
    • A variation of Matej solution might work for prototyping.

Not covered.

Attendees: HamishW, JamesP, JulianO, Matej, Lorenz, Ramon, Thomas Debrunner.

Clone this wiki locally