PX4 Dev Call: November 04, 2020

November 04, 2020

Agenda

  • Status update by component/project
  • Roadmap, and Release discussion
  • Community Q&A
  • In-Depth discussions

Join Meeting


Component update

Hardware Workgroups

Special Interest Groups

  • Korean community dev-call started this week, every Wed at 4 pm KST

System Architecture

It’s possible to disarm using the safety button even if the safety circuit breaker is set.
Take care of:

  • Keep FMU and IO in sync
  • safety button on pixhawk 4 probably FMU and IO
  • safety button is used to boot into bootloader and for other purposes and needs to stay functional

OS / NuttX

NuttX 10 testing is ongoing:

  • USB issue where the MAVLink instance would be preserved when you fly after using the board connected to the computer. This was not always the case.

Driver

New sitance senor driver: @david_s5 's test showed very short maximum range.

Commander

No update

Estimation

Big refactor of optical flow fusion logic. Mostly non-functional changes but also a few bugs were found and fixed. It was tested in SITL and with a bench sensor setup.

We urgently need a test on an optical flow enabled drone.

VTOL

No update

Fixed Wing

@sfuhrer proposes to remove the climb out mode. There is a pr to make it configurable to disable but preferably the mode should be removed since the mode is very easily triggered and does crazy high pitch, high thrust.

Need to review the hand launch case to make sure it’s still working fine.
Action: Change pr to remove it and ping fixed-wing users if defaults need to be changed and if takeoff doesn’t have a problem. Better not have a hidden mode that no one uses or tests.

Multicopter

If there’s no more feedback this will get merged soon:
If there are concerns raise them now.

Rover

No update

Avoidance

No update

Simulation

You can now run PX4 SITL with JSBsim as a separate ROS package.

@Jaeyoung-Lim improved the water level visuals for boat and UUV simulation. Boats are still running on a solid surface. The buoyancy and possibly using a waves model is not moving as fast as the water texture.

MAVSDK

:tada: MAVSDK v0.34.0

  • A lot of MAVLink command supports added
  • Gimbal v2 support
  • Camera improvements
    and much more, see the release notes.

:tada: Python v0.14.0

https://github.com/mavlink/MAVSDK-Python/releases/tag/0.14.0

Updated Python docs:

http://mavsdk-python-docs.s3-website.eu-central-1.amazonaws.com/

:tada: Java v0.5.0 Release
https://github.com/mavlink/MAVSDK-Java/releases/tag/0.5.0

MAVROS / DDS / ROS2

We take part in the ROSWorld 2020. Register if you didn’t, it’s free to join.

https://roscon.ros.org/world/2020/

UAVCAN

Hardware call happening next Tuesday

  • Pulling in new datatypes
  • Testing if serialization is stable and cooperate on bugfixes. There is a branch for improving the serialization
  • DS-015 public document feedback still welcome. It is planned to be released next Tuesday.

The open source drones mini-summit happened. The content will be available on Youtube after a 3 weeks lead time for the paying participants. We’ll share the link.

Release

Fixes on the release branch were not tagged for a point release yet because there were issues reported with the here2 GPS. We’ll probably do a release with the fixes we currently have first.

We need to make sure the original support for here2 GPS needs to be restored or we need to declare it as not supported. Action: Reach out to the community contributor to check in if they have the hardware in use and the same issue.

Idea for next PX4 release workflow is under discussion.

Community Q&A

Kalyan Siriam: Working on a different build system Meson for MAVSDK. He’ll talk to the maintainers Julian and Jonas to make sure it’s well accepted and not a parallel effort

In-Depth discussions

@dagar Improve user experience with devices failing as not being connected even though they are connected but not working and the user should get an error as a result for auto-detection. We should work towards better showing in the UI what devices are connected and support a check against a fixed device tree.

@LorenzMeier Improved configuration storage would help with the device tree specification and with control allocation/mixing. There’s currently draft work being done on this and it would be useful to join efforts to bring it foprward.


Errata and Feedback

Let me know below if we failed to capture anything the right way, and if there are any updates (not present here), or you have feedback you would like to share with the dev-team.

Nothing much to comment on about this, but I’m working on being able to build MAVSDK with the meson build system (currently MAVSDK only supports CMake). This gives users more choice + I personally like meson because it is faster, uses Jinja (a faster build system) by default, and the scripting API is much cleaner. Currently it can only build a subset of the entire MAVSDK project, but hopefully will support the backends and everything else in the near future.