Dev Call Jan 08, 2018 2300 UTC

Development Team
1

Plane

  • 3.8.4 is out

ChibiOS Update

Copter

Rover

  • Rover-3.2 release update
  • Breadcrumb mode

APSync

  • release

Hardware

  • Cube/Pixhawk2 ICM20948 rotation fix

Issues




Pull requests



DroneCore

Funding Report

CoC reminder
New Member Vote

2

Attendee count (max): 24

UTC23:02 Plane and tridge

  • Congrats on new ArduPlane 3.8.4
  • Ms5525 airspeed sensor fix in place
  • Transitions in tailsitters in auto mode fixed
  • Incorrect type of transition
  • Tried to transition like a quadplane which won’t work too well

UTC23:04 ChibiOS

  • Sid and tridge have been working hard on these
  • Successfully flown on PixHawk1 on Iris
  • AP_IOMCU
  • Abstracts IO infrastructure on MCU
  • Uses same firmware on IO processor
  • Chibios-based IO firmware coming!
  • Same protocol initially
  • MdB wants mixers fixed!
  • Mixer upload isn’t functional ATM
  • RAMTron driver (FRAM)
  • Very simple driver
  • Native ArduPilot SPI driver
  • Tiny compared to PX4 equivalent and our compat code for same
  • Byte-addressable memory
  • Parameters will be preserved between PX4 and Chibios so dev is eased
  • You can switch backwards and forwards
  • PR is open against master for ChibiOS
  • https://github.com/ArduPilot/ardupilot/pull/7475
  • Reviews welcome!
  • Testing welcome!
  • Matt has tested on ground for Solo
  • Motors turn…
  • “In thust we trust”
  • Gimbal doesn’t work ATM
  • Oreo LEDs won’t work ATM
  • Also SMBus batteries
  • Missing features are listed on the PR
  • Quite a few
  • But shouldn’t really stop merge
  • https://github.com/tridge/ardupilot/blob/pr-chibios/libraries/AP_HAL_ChibiOS/hwdef/fmuv3/hwdef.dat
  • Used to define a new board
  • Thanks to Jordi and Phil for sending Sid a PixRacer
  • V1 will take a little longer to do because of architectural differences (IO processor via i2c vs uart)
  • Sid’s doing the CCM stuff now
  • 2450GPS port is part of this
  • Same bootloader as normal
  • Could actually write a bootloader in Chibios
  • USB ID comes from apj customisations
  • https://github.com/tridge/ardupilot/blob/pr-chibios/libraries/AP_HAL_ChibiOS/hwdef/fmuv3/hwdef.dat#L20]https://github.com/tridge/ardupilot/blob/pr-chibios/libraries/AP_HAL_ChibiOS/hwdef/fmuv3/hwdef.dat#L20[/url]
  • Matt’s playing with Solo so it understands old and new format USB strings/ids
  • Currently using ST manufacturer ID
  • Peter’s looking at porting to the Spark2.0
  • And the RadioLink PixHawk Mini
  • Should be a matter of bootloader and hwdef.dat for the most part
  • Use same hwdef.dat to create bootloader defines?

UTC23:26 Randy and Copter

  • 3.5.6 coming
  • Fix for Pixhawk2 Cube ICM compass orientation issue
  • Needs someone to test this one
  • Jeff’s volunteered
  • Solo vehicle yawing patches
  • Maybe ADSB min altitude
  • 3.6?
  • Lots of changes in master right now
  • New Loiter still needs to get in

UTC23:30 SmartRTL home patch

  • Rover patch to fix initial home might want to come into Copter as well

UTC23:33 ZigZag PR

  • So easy to add new flightmodes!
  • Use dev flag to mark these things as “experimental”?

UTC23:35 EKF3 to enabled by default

  • Default for Rover and Copter
  • Get this in early, BEFORE we want to do 3.6
  • Plane defaults to ekf3 for tailsitters already
  • Accel learning on all axes is useful for them
  • Generally is OK by tridge
  • 100k of flash per EKF
  • So minimising build requires having a single EKF
  • Moving parameter into AHRS might help things
  • Many #ifdefs required to not have EKF2 or EKF3

UTC23:38 Randy and Rover

UTC23:55 - New APSync for TX1

  • See video
  • OpenKAI (non-GPS navigation)
  • Amilcar’s looking at doing it on the TX2

UTC23:56 - ArduPlane and waypoint injection

  • https://github.com/ArduPilot/ardupilot/issues/7469
  • Implemented in SITL?
  • Injecting waypoints involve shifting
  • append/suspend/resume
  • https://github.com/RUAIAA/ardupilot/commit/bc61c0584afbf50effbd28ff4c42fc9514ddb488
  • OBC has similar issue
  • Points to avoid
  • Dynamic stay-out zones
  • Data coming in over something ADSB-like
  • Continue mission but avoid other aircraft
  • Something like the current avoidance code
  • But make avoidance strategy more complicated
  • Change trajectory rather than running away
  • https://github.com/RUAIAA/ardupilot/commits/obstacle-avoidance
  • GCS makes decisions
  • Discussion on how path planning should be structured
  • Why add waypoints at end of mission rather than using guided to push vehicle around?
  • Fly-via-point?
  • For off-board control
  • Return to original path?
  • On-board avoidance @10Hz
  • Little gridaround aircraft
  • Work out whether you will intersect with other aircraft
  • Postulate new set of points reachable from aircraft position
  • Choose path which gets closest
  • Could use the fly-via
  • Need to tell GCS that avoidance is happening
  • Randy: Simplest change isn’t necessarily the best change

UTC00:19 - https://github.com/ArduPilot/ardupilot/issues/6193

  • Match speed to wpnav_speed immediately rather than at next waypoint
  • Do-change-speed - is that correct way to change vehicle speed?
  • Why set a parameter to change vehicle speed right now?
  • Depends on what you’re trying to achieve?

UTC00:19 - https://github.com/ArduPilot/ardupilot/issues/1348

  • GCS console
  • Nuke from orbit

UTC00:31 - https://github.com/ArduPilot/ardupilot/issues/1348

  • We’ll add a check
  • Parameter which specifies max distance to waypoints
  • Also make sure all mission points are within fence

UTC00:43 - https://github.com/ArduPilot/ardupilot/pull/7429

  • Whitespace-only patches break story and they break PRs
  • Tridge thinks it is up to maintainers as to whether to to accept in their bits

UTC00:52 - https://github.com/ArduPilot/ardupilot/pull/7285]https://github.com/ArduPilot/ardupilot/pull/7285[/url]

  • locking/unlocking port
  • Injecting mga data and the driver itself would conflict
  • The current situation is helpful in the development of drivers
  • Can see the effects in uCenter as you do stuff
  • Tridge has patches around that allow multiple threads to use a uart
  • Use a semaphore rather than checking thread ids

UTC00:59 - https://github.com/ArduPilot/ardupilot/pull/6719

  • head/tail ringbuffers you only end up with 1023
  • Decided not to pursue this PR

UTC12:00 - DroneCore

  • Sid did most mucking around on this
  • Tom will ask him about it
  • Peter’s set_message_interval stuff is relevant
  • Do we even need to support it?
  • Are they even going to continue with DroneCore?

UTC12:04 - Funding committee

  • Closing issues
  • Developed new charter for funding committee
  • Ask for link to document
  • New record on closing a funding request
  • Trying to reach people who made proposals
  • Funding committee helping with partners stuff
  • 23-ish partners
  • Click-and-pledge comes in via emails to ardupilot email account
  • Removal of logs in wiki and website needs to happen
  • Order is important to partners
  • Lapsed partners to drop to bottom?
  • Approval given on call to apologise to partner who didn’t get recognition for 6 months, and to push their re-enrolment date back 6 months

UTC12:18 - code of conduct

UTC12:20 - add Mark and Nick to devteam

  • Tom to create votes

UTC12:21 - contribution of the month

  • Mark WhiteHorne nominated by Luis
  • Tom to create vote for December Contribution of the month

UTC12:25 - drop mode works on side

  • Dropping three at a time
  • Hard to break throw/drop mode

UTC12:26 - unconference

  • 5 weeks to go
  • No shortage of talks
  • Just need to organise them formally!
  • So people can start on their slides
  • We have a projector as we did last time
  • Flight vehicle demos!
  • If people need to ship things they can talk to tridge or Peter
  • Tridge is working on LCA but can focus more after that
  • Hackfest before-hand
  • Randy should be getting largish room on ground floor
  • Francisco will be in mid-Thur, as will Randy, Tom on Tuesday undefined