Dev Call Sept 14 2020

Issues & Pull Requests




1 Like

Attendee count (max): 18

UTC1100 -

  • Bad yaw estimate
  • Copter / tailsitter Plane
  • Bad yaw estimate on EKF3
  • EKF2 is OK
  • Previous test-branch worked OK
  • So something has broken recently
  • Broke about 2 week ago
  • Log shows massive innovations on mag
    • Earth-field estimates are close, ‘though
  • Tridge will look at this off-line with PH

UTC1108 -

  • Remove arming check all option
  • (bit 0 / value 1)
  • Arming force is supposed to replace unsetting the bits
  • But sometimes you need to disable e.g. compass on a permanent basis
  • Maybe another mask - “arm check off”?
  • Use the highest value possible to set the maximum transportable bits
  • Tridge and Peter both like the “inverted” “inhibit check” concept
    • Randy doesn’t

UTC1115 -

  • Old issue that’s been overtaken by patches
  • Issue was actually fixed some time ago
  • Closed

UTC1115 -

  • Merged
  • Travis display improvement

UTC1118 -

  • Moved to DevCallEU

UTC1120 -

  • Probably want to do a bit of refactoring around this area
    • Move it into AP_Vehicle
    • Remove printfs directly
  • Send-text in ap-vehicle then this can be merged

UTC1125 -

  • Merged!

UTC1126 -

  • Blasting all of the values isn’t great as some of them are carefully crafted
  • Too broad a change
    • Thought needs to be put into each individual change

UTC1131 -

  • If a goto command is rejected then the vehicle keeps doing what it’s doing
    • E.g. fence in the way
  • Could instead fly to point at edge of fence?
  • Merged!
  • Related
    • Add more checks around Missions to make sure they’re sane

UTC1136 -

  • Lock in versions of gcc which don’t produce any warnings on stm32 builds
  • Two whitelisted compilers ATM
  • 4.9.3 and 6.3.1
  • Only applies to C++ code
  • C code produces some warnings
    • H7 processors on ChibiOS
  • Whitelisting in SITL might come along
  • Merging this might break master
    • Depends on what’s been merged today

UTC1141 -

  • Really expensive message
  • Worried about bandwidth
  • Can decimate onboard producers
  • Parameter AVD_OPTIONS to enable the message?
  • Maybe a way to decide which message to use?
  • Choose between rangefinder distance and obstacle distance depending on bandwidth?
    • Probably too confusing

UTC1155 -

  • Randy needs to review this

UTC1156 -

  • Definitely need to avoid using existing logging system for logging EKF inputs
  • Merged

UTC0002 -

  • Correct change
  • Real issue
  • Hasn’t checked the implementation
  • Michael wants ot check with uCenter
  • Can’t allocate lots and lots of channels to Galileo
    • Wouldn’t allocate any so wouldn’t track Galileo
  • MdB will merge when ready
    • Having trouble finding time
  • Tridge might also find time to test
    • Gnss mode 71
    • Symptom?
      • Won’t get satellites

UTC0004 -

  • Revisiting an old one
  • Bounce to Andy?
  • Power_watts could be incorrect in some cases
    • E.g. fuel-based systems
  • Should we have a batt-index instead of a sum?
  • Getting rid of places which nominate 0 or use the default battery number
    • Places will need to nominate their battery

UTC0014 -

  • 4.0.6 is out
  • Will be a 4.0.7
    • Compass failover problem
    • Currently active compass dies completely (e.g. canbus cable comes loose)
      • EKF will not fail over to second compass
    • Works in master
    • Worked in early 3.6 Copter releases
    • Broken for a long time
    • There is some observability of yaw when moving along in a Copter
  • No nasty reports from 4.0.6
  • 4.1 beta period within next month
    • GSF is a big drawcard
  • Airspeed affinity was flown on the weekend
    • Worked as designed
    • Dual-airspeed is a real win as failed airspeed sensors has been a Plane-killer

UTC0018 -

  • For 4.1
    • Randy would like to get EKF sensor source stuff
    • S-curve navigation
      • Back on top of Randy’s list
      • Another week before a PR comes in
      • A solid month and Copter might be ready
  • Still no Copter stable release
    • MP prioritisation table problem
    • Hopefully MO will get to this today

UTC0023 - Rover?

  • Zero-turn-mower
    • Two F9s for yaw
    • Struggle to set up the GPSs
      • Need to look at why he struggled so much

UTC0024 - User Alerts

  • Just about ready to roll!
  • Documented on Wiki
  • ArduPilot repo is up and running
  • Do we want to put issues prior to now into the database?
    • Disadvantages
      • Accuracy?
      • Completeness?
    • Advantages
      • Quick win for those using older firmwares
  • Candidates
    • CubeOrange parameter resets
    • IMU lockup problem when hardware goes away (the one that made us do watchdog etc)
  • Will let us test the process of adding user alerts as well
  • Tridge will work with Stephen to get an entry in
  • MAVProxy already has an implementation
  • MP is getting patches
  • Blog post coming explaining what the DB is and how we’re using it

UTC0034 -

  • Need a junkcode test to make sure subtracting a pair of uint16_t’s does the right thing

UTC0042 - vibration compensation?

  • Heuristics are relatively easy to find
  • Need a tlog

UTC0045 - Mower log

  • Very nice yaw from the GSF!
  • Mowers + GSF might be a thing!
  • Primary GPS keeps dropping offline

UTC0052 - Closed