Dev call Nov 27, 2024

CraigElder · November 25, 2024, 10:58pm

Copter, Plane, Rover Update

ArduPilot 4.6 issues list

opened 12:01AM - 14 Nov 24 UTC

DevCallTopic ReleaseAdmin

This is a list of issues discovered during ArduPilot 4.6 testing. Issues for [C…opter/Rover-4.5 issues are here](https://github.com/ArduPilot/ardupilot/issues/26103) **Reports requiring investigation** - [ ] Plane takeoff not completing under TKOFF_OPTIONS[0]=0 #28685 ([PR](https://github.com/ArduPilot/ardupilot/pull/28707)) - [ ] Custom Build Server for CubeOrangePlus with OpenDroneID shows linker error ([discussion1](https://discuss.ardupilot.org/t/copter-4-6-0-beta1-is-available-for-beta-testing/126295/4), [discussion2](https://discuss.ardupilot.org/t/copter-custom-firmware-build-issue-with-remote-id-support-latest-not-4-5-7/125177)) **Confirmed issues** - [ ] memory fragmentation and multiple memory regions mean lua scripting can fail with out of memory error even if you have plenty of free memory. Addressed in https://github.com/ArduPilot/ardupilot/pull/26517 - [ ] system time may be unstable when using only first GPS timestamp, consider https://github.com/ArduPilot/ardupilot/pull/24022 or similar fix (Tridge has details) - [ ] on ICM45686 IMUs the SPI transfer overheads can be overwhelming on some boards, especially those with 3 of that sensor. Some PRs to address: https://github.com/ArduPilot/ardupilot/pull/28672 and https://github.com/ArduPilot/ardupilot/pull/27573 ***Plane Specific Issues*** - [ ] plane takeoff regression since 4.5.x: https://github.com/ArduPilot/ardupilot/issues/28681 - [ ] on plane when FENCE_AUTOENABLE=3 and FENCE_ENABLE=0 the fence does not enable when arming if there is a RCn_OPTION set to FENCE. This is a regression from 4.5.x. (Tridge has details) Fix: https://github.com/ArduPilot/ardupilot/pull/28729 - [ ] on plane the pre-arm check for being inside polygon fence was lost since 4.5.x. Fix: https://github.com/ArduPilot/ardupilot/pull/28729 **Enhancements or Issues that may not be resolved before the stable release** - [ ] Add LDRobot LD19 support ([discussion](https://discuss.ardupilot.org/t/copter-4-6-0-beta1-is-available-for-beta-testing/126295/8), [issue](https://github.com/ArduPilot/ardupilot/issues/28662)) - [ ] need to provide a clearer message to users when the power status flags change, indicating loss of a power source in flight. A statustext at message level CRITICAL I think is warranted, maybe "POWER STATUS CHANGE" ? (Tridge has details) **Wiki issues/enhancement requests ([Wiki To-Do list](https://github.com/ArduPilot/ardupilot_wiki/issues/4274))** - None **Resolved issues** - [x] Plane fences ignored ([issue](https://github.com/ArduPilot/ardupilot/issues/28666)) -- cannot reproduce

Issues & Pull Requests

github.com/ArduPilot/ardupilot

Tools: scripts: Update MAVLink parsing for wiki inclusion

ArduPilot:master ← ES-Alexander:mavlink-messages-rst

opened 09:40AM - 03 Jun 24 UTC

ES-Alexander

+88 -5

ArduPilot Wiki uses reStructuredText (rst) source files, so supporting rst outpu…t is essential for this script to generate wiki files. Ideally this can be run automatically for every vehicle type and version, like the existing parameter lists. - [x] Add rst support - Example file for ArduSub 4.1, with all options included: [ArduSub_Sub-4.1_MAVLink_Messages.rst.zip](https://github.com/user-attachments/files/15531781/ArduSub_Sub-4.1_MAVLink_Messages.rst.zip) - **EDIT:** Some additional example files from the `master` branch, for different vehicle types (all message options included): [MultiVehicle_master_MAVLink_Messages.zip](https://github.com/user-attachments/files/15793798/MultiVehicle_master_MAVLink_Messages.zip) - **EDIT2:** Example files for released firmware versions have been [added here](https://github.com/ArduPilot/ParameterRepository) --- **EDIT3:** Wiki inclusion and future planning notes moved to ArduPilot/ardupilot_wiki#6065

github.com/ArduPilot/ardupilot

Quicktune C++ conversion (WIP)

ArduPilot:master ← MichelleRos:pr/quicktunecpp

opened 07:08AM - 17 Jul 24 UTC

MichelleRos

+1024 -11

Add Quicktune to C++ code (conversion from the VTOL-quicktune lua script). Te…sted in SITL, multiple multcopters and quadplanes Edit: Tridge has added support for quadplanes and an autotest for it and made it not run for helis. I plan to add an autotest for the copter version shortly. I have also updated the PR to limit quicktune to only alt_hold and loiter in copter, and Tridge did something similar for quadplanes. Edit 2: Tridge updated it to make it always allocate, to be more in line with the pattern in copter, and moved it to its own thread. He also added a max attitude error, so it auto-aborts if the vehicle starts oscillating badly during the tune and made it abort the tune if the pilot changes to a mode that doesn't support quicktune. Note that it does not abort if the pilot changes between modes that do support quicktune, allowing people to, for example, do the tuning in loiter mode, then switch to alt_hold to test the tune before saving the gains. Now tested on a quadplane as well. As for the flash concerns, yes, it uses around 5k of flash, but autotune uses around 15k of flash, so for a tuner, quicktune is actually rather light.

github.com/ArduPilot/ardupilot

Copter: allow adding custom modes from scripting

ArduPilot:master ← IamPete1:copter_custom_modes

opened 12:26AM - 15 Sep 24 UTC

IamPete1

+516 -65

This allows a script to register a new mode with a given number and name. This i…s just a wrapper on guided mode, but it allows clearer interaction with the user. The example re-implements flip mode in a script. It uses mode 100, in MAVProxy you can enter with `mode 100`. In order to allow the flip I added a new guided binding `set_target_rate_and_throttle`. I also found and fixed a issue that the Z controller would throw a flow of control error when moving from guided angle with throttle to guided angle with climb rate. The internal variables of guided mode are changed to static so they apply to to the new modes too. There are some more changes needed: - [x] anywhere that doing a compare on guided mode by reference or number needs to deal with the new modes. - [x] Come up with a way to communicate to the GCS which modes are available so it could populate its mode list with the new modes (this would also be useful if we want to start removing modes and/or for the custom build server). - [ ] Currently scripting is the only way to register a mode, but there is no reason a companion computer could not do the same with a new MAVLink message. - [x] In the future we may want to add more state on the custom mode object, for example a "allow_enter" bool which the script could keep updated. This would allow the script to reject entry if the correct conditions are not met. Currently the flip example will enter but then immediately put the mode back again if it does not like it. - [ ] Allow a custom version of guided no-gps as well as full guided. - [x] Currently the mode will fail to add if the script is restated because it is already there, it might be nice to allow a re-add of the same mode, although this is only a issue for script dev.

github.com/ArduPilot/ardupilot

AP_Mission: `DO_RETURN_PATH_START` limit worst case runtime

ArduPilot:master ← IamPete1:mission_return_path_worst_case_runtime_limit

opened 07:44PM - 13 Nov 24 UTC

IamPete1

+12 -6

If you fill your mission with `DO_RETURN_PATH_START` way-points the best rejoin …location calculation will take so long that the flight controller watchdogs. This adds a worst case limit of 1000 search points. The limit used to be 254 points per `DO_RETURN_PATH_START`, the the new limit is equivalent to four maximum length return path options. This change allows a single path to be longer. https://github.com/ArduPilot/ardupilot/pull/28574

github.com/ArduPilot/ardupilot

AP_NavEKF3: pos and vel resets default to user defined source

ArduPilot:master ← rmackay9:ekf-reset-source-fix

opened 01:04AM - 21 Nov 24 UTC

rmackay9

+40 -0

This resolves issue https://github.com/ArduPilot/ardupilot/issues/27729 by using… the user defined position or velocity source (e.g. EK3_SRCx_POSXY, EK3_SRCx_VELXY) whenever the caller does not specify the source. This solution is a minimal-change solution to the problem. Another possible solution would be to remove ResetVelocity and ResetPosition's xxxResetSoiurce argument completely and instead always just use the user defined source. This has been lightly tested in SITL to confirm that it resolves the issue. Below are before and after screen shots of SITL testing where the Vicon has been turned 90degrees from reality. Before we see terrible behaviour (even worse than described in [the issue](https://github.com/ArduPilot/ardupilot/issues/27729 )) where the lane switches every 7 seconds and the vehicle resets to the GPS location. This results in the vehicle flying an ever expanding circle around the target at full speed ![before-map-and-console](https://github.com/user-attachments/assets/2cd2d311-3a45-496f-b3dc-2642670f1874) After we see nearly twitch-less transitions between Optical flow and External Nav. The two vehicles shown are the EKF's estimated position and the actual position and their positions are stable ![after-map-and-console](https://github.com/user-attachments/assets/98068474-2ce8-49ca-9b41-39f1a16a11d8)

CraigElder · November 26, 2024, 4:14am

Attendee count (max): 18

UTC2300 - too far?

Too few PRs marked for DevCall?

UTC2301 - Plane update

Tridge has a PR for the takeoff regression
- ArduPilot 4.6 issues list · Issue #28612 · ArduPilot/ardupilot · GitHub
- Fixed takeoff direction with no yaw source
Fence stuff
- Two regressions so far on 4.6
- Despite a vast amount of testing
- AC_Fence: fixed 2 regressions from 4.5.x by tridge · Pull Request #28729 · ArduPilot/ardupilot · GitHub
  - Change test or change fix?
Plane takeoff issue
- Close to altitude then hold and never reached
- George has a fix
  - Start to use TECS in last part of takeoff
  - Discarded concept of integrator in takeoff code
    - would probably end up with a jerk at end of takeoff
    - And is basically reimplementing TECS
  - Plane: Fix inability to climb the last few meters of takeoff by Georacer · Pull Request #28707 · ArduPilot/ardupilot · GitHub
- probably a regression 4.5 to 4.6

UTC2313 - Copter update

Scarily low response rate to call-for-beta-testers
- Several saying everything’s OK
- We need beta testers! Please test!
  - Report success as well as failure….
- Devs as well as end-users
CubeOrange / OpenDroneID build problem in Custom build Server
LD19 Llidar support requested
- Rishabh might do it

UTC2317 - Tools: scripts: Update MAVLink parsing for wiki inclusion by ES-Alexander · Pull Request #27226 · ArduPilot/ardupilot · GitHub

Mavlink parsing script
Generates rst for mavlink messages
- Not including minimal.xml
- So might be missing heartbeats?
Probably just need to use “all” dialect

UTC2326 - https://github.com/ArduPilot/ardupilot/pull/27578

Quicktune C++ conversion
PH thinks its pretty good
Peter will have a look at this after the call and merge it if he’s happy

UTC2330 - Copter: allow adding custom modes from scripting by IamPete1 · Pull Request #28110 · ArduPilot/ardupilot · GitHub

Allow adding custom modes from scripting
Ap objects can have fields in them
- Allow-enter
Never remove modes
- Scripting restart will get you the same object back!
Probably want to add to autotest suite
- LUA flip-mode
- Use freestyle quad
PH will be back with this one

UTC2344 - AP_Scripting: add Plane autoland applet by Hwurzburg · Pull Request #28591 · ArduPilot/ardupilot · GitHub

Plane autoland applet
Parameter for reverse-course or same-course landing?
Need an autotest for this
- So the applet keeps working
Takeoff code knowing whether the yaw is initialised
Tridge is working with Henry on this one

UTC2356 - AP_Mission: `DO_RETURN_PATH_START` limit worst case runtime by IamPete1 · Pull Request #28608 · ArduPilot/ardupilot · GitHub

Limit worst-case runtime for DO_RETURN_PATH_START
Nice to put comments on timing measurements made
Was possible to watchdog…
- But even just waiting 1.8 seconds between motor updates is bad!
IO task callback which precomputes?
- Assume your decision won’t change that often?
Prearm check for a landing waypoint if you have a do-return-path?
- Leonard points out you might not want a landing point at the end
Long-loops might also be a problem for DO_LAND_START
Are we preventing valid missions working?
This PR isn’t ideal, but better than what we have…
merged!

UTC0116 - AP_NavEKF3: pos and vel resets default to user defined source by rmackay9 · Pull Request #28693 · ArduPilot/ardupilot · GitHub

Pos and vel resets default to user-defines source
We reset to the wrong location!
- Eg. to GPS when your sources don’t specify it to be used
This doesn’t fix every possible problem
Should we switch to using a median solution from the EKF in case of three lanes?
- Problematic when the three IMUs are not comparable
  - Eg. two isolated, one non-isolated
We need more documentation for indoor/outdoor navigation
- Consider locking to a single lane?
Randy will come back to this and consider just resetting to the current source
- And not resetting if you don’t have one
Generally need to come back and think about our lane-switching logic
- Core-error-score stuff isn’t really working
- Lane-scheduling, single significant baro glitch can cause lane change, even if it will be consumed by the next core in the next loop

UTC0138 - open vote

Devs please check

UTC0138 - close

UTC0139 - brown-out protection on H743?

Is it a good thing?
- If you’re flying you might reset if the voltage drops
  - Almost guaranteed to crash the vehicle
- If you don’t then you’ll probably survive
- Would be nice to be notified that you would have browned-out
Some systems effectively do it
- Eg. standard CubeOrange setup
- Measured at a different level
  - Supply voltage
  - Not MCU voltage
Low-powered vehicle
- Some sort of a sleep mode?
- External relays work….
Issue raised