Dev Call Mar 25 2019 2300 UTC

Developer Conference Update

Issues & Pull Requests

Antenna Tracker




Attendee count (max): 16

UTC2304 - Conference

  • Very successful, a lot of fun

  • Luis is working on the videos

  • Lost audio on some lightning talks

  • Post-conference fun, too

    • Plane built
  • Many thanks to Grant and James and Daryl for bbqs and the like

  • ~45 attendees on Saturday/Sunday

  • Leonard’s ArduKid t-shirts

    • Birthday cake too!
  • Night flying was good

  • Demos on Thursdays

    • Whole host of problems on new boat

      • Forgot to plug hole in back of boat

      • Badly tuned

      • Got most of it sorted

    • Got things sorted on subsequent visits

      • And have video including pov
  • Iain / blurobotics was great!

    • Both in lake and in hotel pool
  • Scripting was good

    • Long talk

    • Demo

      • Aion Robotics and Peter’s rover

      • hotter/colder game

  • Swift QuadPlane was awesome

    • Very refined

    • Flies beautifully

    • Attention to detail

    • Demonstrated new sort of landing

      • Landing into the wind

      • Big circles required

        • Had to walk out a ways
  • Tandem-heli / Rover

    • didn’t fly

    • Heli + Rover + fire extinguisher

    • Fire-fighting robot

  • Videos in next day or so

UTC2314 -

  • Merged!

UTC2314 -

  • Need Jaxxzer to say OK but everyone else is OK with it

UTC2326 -

  • Merged

  • Header cleanups

UTC2328 -

  • Only used by attitudes_consistent

    • And get secondary quaternion

      • Only used for logging
  • Prearm checks were failing for Mark on tailsitter

  • Will change logging of quaternion

    • May affect Paul

    • Could localise to attitudes-consistent to avoid logging changes

  • Needs Paul to OK it

    • Added as reviewer

UTC2332 -

  • Makes it easier to test/fly body-frame roll option

  • Needs a rebase

  • Two different control modes

    • If you are flying multi then you expect roll to operate in earth frame for Copter

    • Body-frame-yaw if pointing vertically in tailsitter

  • _M is not good

  • Can go in after a rebase

UTC2339 -

  • Q_trim_pitch was briefly broken

  • Plane: VTOL Pitch trim bug fix for non tailsitters

  • Does a transpose when first setting trim rather than at runtime

  • Fixes mavlink

UTC2341 -

  • Merged!

  • We really should broaden compiler support

UTC2343 -

  • No objections on devcall

  • Peter to work with Pierre on getting this in

UTC2350 -

  • Randy likes it

  • Follow-up PR

  • Need to add checks for heli

  • Peter to add #if guards around check and can then merge it

UTC2359 - crow flaps

  • Need tridge to look at it

UTC0000 -

  • Merged

  • Can sort problems out

UTC0005 -

  • Discussions on why this is failing in CI

  • Randy to test on his machine

UTC0006 -

  • Spool logic patch has to go in before this can go in

UTC0009 -

  • Stuck behind spool logic patch

UTC0015 - tracker release

  • Peter Hall to split fix-scan mode out

  • Start building Tracker for all boards

UTC0022 - Plane update

  • Onion has been flown on two planes

UTC0023 - Copter update

  • Number of people asking cxof optical flow sensor available on cheaper boards

  • Openmv as optical flow camera is a thing

    • There’s a script that runs on the camera and sends mavlink messages
  • Beta will come along sometime….

UTC0024 - dronekit and ArduPilot

  • Bring it under umbrella of ArduPilot

  • Anything not on chat?

  • Discussion of how transfer would happen

  • Vote required

  • My apologies for the rambling nature of this post. I’m going to get another coffee. OK, my 2c on the dronekit-android and dronekit-python thing. I can really only speak for dronekit-python here. These are APIs meant to abstract mavlink into a vehicle model which a Python/java programmer can use to manipulate the vehicle. That abstraction is important. A lot of the most recent contributions to dronekit-python have come from PX4 users - and I have bounced patches that moved the API to be ArduPilot-specific. There’s been NO talk in here of talking to the relevant communities (such as they are) to see what they think of the concept of ArduPilot taking over these repositories? Will non-ArduPilot users drop dronekit-python, believing they’ll get screwed over? We just had a conference where we discussed what we should take on or drop as a project - bizarrely that didn’t even cover taking over dronekit-python! Why would be ever drop anything as a project if we’re willing to pick up a project and then apparently be happy to do absolutely nothing with it - where the act of taking it might actively harm that project? Promises of future support for the project where there has been none also count for very little in my eyes when making a decision like this; nobody has really associated dronekit with 3DR for a long time AFAICS.

  • Luis should we be stretching ourselves like this?

    • Won’t it look stale?

    • Randy: do we have the resources in the devteam to maintain this?

    • Luis: does it make sense for ArduPilot to maintain DroneKit?

      • Unless we have people commited to develop dronekit we shouldn’t take it on
  • Resourcing issue

    • Who is going to maintain it?

      • Kelly has his own fork

        • Someone has only asked him if he will take it on
    • This is a serious impediment to take this on

    • Don’t just try to volunteer Peter’s time

  • Randy: Need to decide other issues BEFORE we take on the project

    • Don’t want to take something on and let it die in our repo
  • Is this aligned with ArduPilot as a project

    • Randy thinks “yes!”

      • But resources….
  • Why move things under the ardupilot project rather than just take dronekit on?

  • This sort of thing happened before

    • E.g. Tower
  • Randy we do want an SDK and dronekit is the closest thing to it

  • Luis: Why did PX4 give up on it?

    • Randy: don’t care - what do we want?
  • Any vote has to cover these points

    • E.g. “should we take this on with no resources allocated?”

    • E.g. “should we take this on only if resources are allocated?”

UTC0045 - first video of conference will be up in a few minutes

  • Wireless mic was not good