PX4 Dev Call: May 13, 2020

šŸ“ž Coordination Calls PX4 Autopilot Coordination
1

May 13, 2020

Agenda

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

Join Meeting

Meeting ID: 946 175 205
Join using your mobile/desktop

Call in using your phone, find your local number: https://zoom.us/u/aetSYKbiMF

PX4 Virtual Developer Summit July 6 & 7

New website live: https://px4.io/virtual-2020/

Everyone in the community has the opportunity to have a 30 minutes slot for a talk or video. Get in touch and we might even adjust the schedule if it doesnā€™t fit your content.

Sales pitches for your product are not welcome. We are looking for technical content to help the community cooperate on the open-source projects.

Presentations can be in your native language and there will be regional content e.g. in China.

Ideas for content:

  • Community tour: All the platforms, tools, places to find information and work together. Easy entry.
  • How to set up: Software and Hardware how to get started.
  • Hands-on tuning guide.
  • Presentation of specific components.

There will be a real developer summit planned again when the global health situation is over.

Component update

Flight Testing

Still no official test team. We are trying to get a replacement soon.

System Architecture

All messages that are sent to the ground

OS / NuttX

@david_s5 NuttX made a new release 9.0.0, we didnā€™t test it yet.

Driver

Bosch sensors BMI55/88 tweaks to only publish valid data and have gyro and accel data synchronized.

There was a bug of the estimated gyro bias getting subtracted twice from the measurement.

Commander

No updates

Estimation

From @kamilritz: Enables usage of body position reference frame for VIO usage.

Our release blocking issue seems to be solved on master. Weā€™re not exactly sure what made the difference.

MAVLink

This needs to be reverted. Itā€™s deprecating frame types that we use and are replacing already deprecated other ones.

VTOL

Fixed Wing

Issue from release testing. Rudder wheel track

Multicopter

Workaround for acceleration feed-forward interfering with existing takeoff ramp logic.

Handle estimator attitude resets in between estimation change and new setpoint from position controller:

Optional setpoint filtering for Manual/Stabilized mode:

MAVSDK

C++ autogeneration is finished. @TSC21 contributed a feature for calibration.

Releases!
MAVSDK v0.26.0
MAVSDK v0.25.0
MAVSDK-Java v0.3.1
MAVSDK-Python coming tomorrow

Make sure to check out the release notes which will greatly simplify adaption work to new versions!

MAVROS / RTPS / ROS2

No updates

Avoidance

No updates

Simulation

MAVSDK CI test failure rate in CI is quite high.

  • @JulianOes was looking into it. There seem to be connection issues to simulation that he couldnā€™t reproduce locally. It could because by the dynamic multiple world spawning change.

  • @kamilritz Knows about the ā€œground truthā€ issue where the EKF sets local position true just at the beginning and if it takes off with the false flag

  • @Jaeyoung-Lim : sitl_gazebo update results in a failure. It seems to be only offboard attitude control. Strange because the changes should not change the result. We need to check

UAVCAN

  • @TSC21 The DS-015 UAVCAN Drone Standard RFC should be closed this week so the respective DSDL files can be written to mirror the messages on the spec.
  • Progress by @dagar in the UAVCAN v1 integration in PX4 can be followed in:

e.g. bridging certain messages like battery status between UAVCAN v1 and uORB

Thereā€™s a drone specific UAVCAN weekly call. ā€œUAVCAN Droneā€ on Tuesday in https://www.dronecode.org/calendar/. Any cooperation and help is appreciated. Thereā€™s a general (non donre specific) UAVCAN dev call on Wednesday. Thereā€™s a lot of progress at the moment.

Roadmap, and Release discussion

PX4 v1.11.0-beta2 is out

Post your test flight logs on: PX4 1.11 Beta is here! Community Testing Week Starts Today!

Please help us testing and make sure everything you care about works as expected in the release. If you donā€™t know what to test there are prepared test cards available here: https://dev.px4.io/master/en/test_and_ci/test_flights.html

Community Q&A

Noone raised a question.

In-Depth discussions

Gliding / Airspeed prioritization for fixed wing flight

@Jaeyoung-Lim We want to have fixed wing gliding and soaring control be available from offboard. Instead of making a new setpoint type the idea is to have a special e.g. negative mission throttle interpreted as command to glide so TECS prioritizes airspeed over holding altitude. Is that viable? @dagar We only need to check if any existing functionality interferes.

@sfuhrer How about enabling the feature for RC controlled flight? It would be useful to have a mode where itā€™s impossible to stall the vehicle but have manual control over throttle. @dagar wanted to enable that a long time ago since some commercial planes offer that as the lowest level operator proof mode.

@Jaeyoung-Lim The airplane in SITL does not stall. Default simulation environment will need to change for fixed-wing. How about using JSBSim as the headless simulator and use FlightGear as the visualization. SITL gazebo for fixed wing requires a lot of manual improvements since the lift/drag model does not account for stalling, zero angle of attack results in zero dragā€¦ It will not easily have good dynamics.

Errata and Feedback

Let me know below if I failed to capture anything the right way, and if there are any updates, or you have feedback on the call format.