Dev Call Mar 11 2019 2300 UTC

Development Team
1

Issues & Pull Requests

ROS

General

  • SD card reliability
  • Parameter reset issues
  • SPI Bylaw changes

Google Summer of Code Google Summer of Docs

We’re delighted to announce the inaugural year of Season of Docs, a Google program that fosters collaboration between open source projects and technical writers. Season of Docs is similar to Summer of Code, but with a focus on open source documentation and technical writers. Details are on our website: g.co/seasonofdocs.

Would you like to take part as an open source mentor in the inaugural year of Season of Docs? Open source organizations can start thinking now about the projects you’d like a technical writer to work on. Take a look at the examples of project ideas. Reach out to your community members to see who’d like to be a mentor for Season of Docs. As a mentor, you don’t need technical writing skills. Instead, you’re a member of the open source organization who knows the value of good documentation and who is experienced in open source processes and tools. See the guidelines on working with a technical writer. Organization applications open on April 2, 2019. See the full timeline on the Season of Docs website.

From April 30, Technical writers can explore the list of participating open source organizations and their project ideas. Technical writers bring their skills in designing and developing documentation to the open source organization. Technical writer applications open on April 30. The list of accepted technical writing projects is announced on July 30.

Please do tweet and blog about Season of Docs if you’d like to share the news. We want as many people to know about it as possible. We’ve provided logos that you can download and some example content on the press page.

We’re looking forward to an exciting pilot of the Season of Docs program!

Plane

Copter

Rover

2

Attendee count (max): 17

UTC2302 - https://github.com/ArduPilot/ardupilot/pull/10724

  • Moving longitude scale

  • merged!

UTC2311 - https://github.com/ArduPilot/ardupilot/pull/10708

  • Adding rudder to tailsitter scaling gains

  • Straight-forward to tridge

  • Merged!

UTC2311 - https://github.com/ArduPilot/ardupilot/pull/10737

  • Class on enumerations in arming

  • Use AP_Arming::Method rather than Method then can be merged

UTC2312 - https://github.com/ArduPilot/ardupilot/pull/10703

  • Use Controller pre-arm checks in QuadPlane

  • Bulk of patch is just making sure they work

  • merged!

UTC2313 - https://github.com/ArduPilot/ardupilot/pull/10699

  • If motor output is inhibited in a q-mode then you are not flying

  • Side effect of being able to in-flight tune a compass on the ground

    • New compass-learn method of four

    • But, but - the throttle check?!

    • [10:17 AM] (Channel) MdB: But the is_flying loop hole should disappear IMO :slight_smile:

  • May cause knock-on effects

  • Mark will look at interaction with QACRO patch

  • merged

UTC2323 - https://github.com/ArduPilot/ardupilot/pull/10783

  • Always check return values for origin functions

  • Good idea?

  • Just one spot to fix in this case

  • Get_position is a little odd

  • Several more accessors could do with the same check

  • merged

UTC2329 - https://github.com/ArduPilot/ardupilot/pull/10686

  • Autotest for throw mode

  • Are we OK changing the SIM_ parameter behaviour

  • Yes - this can be merged

  • So long as test suite passes with SITL parameters they’re OK to change so long as you don’t do anything silly

UTC2340 - https://github.com/ArduPilot/ardupilot/pull/7287

  • Date-in-build

  • De-devcalled for the time being

UTC2340 - ensure compass is healthy before using it

  • https://github.com/ArduPilot/ardupilot/pull/6912

  • FF found a problem

    • MdB is finding ways around the problem

  • Adds more ways to not do compass switching

  • Tridge would like to sort compasses putting external first

    • Including canbus detected devices

    • Separate PR

UTC2343 - https://github.com/ArduPilot/ardupilot/issues/10650

  • GPS week roll-over event

  • MdB thinks we’re fine, just close it

  • Devcalltopic removed

UTC2344 - https://github.com/ArduPilot/ardupilot/pull/10614

  • Crow-flap parameters with Mr Hall

  • No updates, waiting for tridge to play with it on RealFlight

  • Model is up

    • Hand-launch?!

    • Change maps to one using bungee launch

    • Can’t see control surfaces on the ground

      • Hand-launch into ground (crash then look)

    • May be a variant with a motor

    • Tridge is trying to get a collection of virtual vehicles together for testing

UTC2351 - request_message

  • Can be merged once it passes CI

UTC2352 - https://github.com/ArduPilot/ardupilot/pull/10406

  • AP_AHRS: apply AHRS trim in EKF_TYPE_SITL

  • Tridge notes it works OK in realflight and sitl

  • merged!

UTC2352 - https://github.com/ArduPilot/ardupilot/pull/10391

  • Storage rearrangements

  • Tridge would like testing done to make sure that missions and mavlink2 keys work

  • Set up signing before/after applying these patches

  • Some boards have extra storage areas

    • More space for parameters

    • Need to make sure OSD data is preserved

      • Tridge will need to review and test on real hardware for this

UTC2356 - https://github.com/ArduPilot/ardupilot/pull/8587

  • Support for extended-sys-state

  • Quadplane stuff has been split off

  • Sub hasn’t been attempted - can there be a mapping for Sub for this message?

  • Some reports of MAVlink bandwidth getting too high

Discussion of issues with Dragon Link - only 19.2 kBaud

V3 is faster

  • Use of message intervals will help

  • Also low-bandwidth-mavlink

  • Discussion on bandwidth requirements and instance messages

    • Randy thinks if someone asks for IMU data at 5Hz then they should get each of the sensors each at 5Hz, not increment through each of the sensors at 5Hz

UTC2359 - https://github.com/mavlink/mavros/pull/1184

  • Fancy Rover off-board control with ROS

  • Set-position-target

  • Off-board path-planning

  • Guided-enable and guided-limits commands supported

  • In auto

    • Set-position-target-lobal-int is sent to ROS path planner

    • Sends back velocity commands which we follow

  • Still have several configuration issues on the ROS side

  • ArduPilot sending stuff to ROS

    • MAVROS guys not-so-happy with that approach

    • Want a new mavlink message

      • Specs don’t seem to indicate this is required

  • Rover Object Avoidance using ROS for Path Planning

UTC0012 - SD card reset

  • Parameters resetting

  • Resolved in master - hopefully

    • Please tell tridge if you get parameter resets on master

  • Reliability on SD card writes and flash operations have gone in

  • ChibiOS gitter channel

UTC0013 - SPI vote

UTC0015 - Google Summer of Docs

  • We need to promote this and encourage people to do this

  • Need to put an application in for people to do google docs

  • Exposure to more technical writers, perhaps?

  • No impact on GSoC

    • Except spreading our resources

UTC0016 - tridge and Plane

  • New point release soon

    • Particularly the retries on SD card

    • QACRO is in, of course

      • Being flown by lots of things on the forums

UTC0018 - verbose is broken

  • “waf all” with --verbose is broken

    • But not with ultra-verbose?

  • Issue to be opened

  • You can use gdb to attach to a Python instance and get a python line number which it is running through at your breakpoint!

UTC0019 - https://github.com/ArduPilot/ardupilot/issues/9498

  • Randy and Copter update

  • Vibrations are still a problem on some vehicles

    • Maybe look at the change in altitude vs climbrate

    • Can set weighting of baros vs accels

      • Maybe push this to relying more on baro

        • Didn’t work on SkyViper

          • Had some elaborate patches to fix

            • Inertialnav and dcm

            • Flying dcm and inertialnav

            • Flies quite smoothly

              • Better than EKF

                • EKF would work most of the time

                  • Diverged

            • Could bring these back into master

              • Fall back to this on EKF failsafe?

            • One other option is to cap throttle at ⅔ hover throttle instead if you get a failsafe

          • Few reports of problems after the dcm/inav fixes

    • Just using baro causes problems as the baro is laggy

      • Oscillations

    • Another option - don’t use z-element of accelerometers

      • Use length-of-gravity + x/y components

      • Used this way back when with ArduPilot

      • Falls apart when you have high turn rates

    • Our AHRS interface is inconsistent

      • On DCM some things returns false

      • This changes it so these functions return true because the data is available

  • 3.6.8 beta release coming soon

    • Fix rare issue with GPS blending

    • And SD card fixes

UTC0033 - https://github.com/ArduPilot/ardupilot/issues/10350

  • Nothing to report for Rover

  • Growing issue of fixes that need to go out