April 01, 2020
- Status update by component/project
- Roadmap, and Release discussion
- Community Q&A
- In-Depth discussions
Meeting ID: 946 175 205
Join using your mobile/desktop
Call in using your phone, find your local number: https://zoom.us/u/aetSYKbiMF
- New docs for compass power compensation
Discussion about Navigator, flight modes improvements: https://docs.google.com/document/d/1N59JHFqL_E9X4b-U0CzxW_eydqXUHZMxAH_FTD30ilk
Anyone interested: We’re open to discuss and coordinate on the PX4 slack channel #architecture.
OS / NuttX
@david_s5 Is testing SPI DMA on H7 e.g. Holybro Durandal and new CUAV boards to save 30% of CPU. But it doesn’t work yet. It would be a nice to have while switching to the new IMU drivers.
@david_s5 Is looking at NuttX upstream changes and saw semaphore changes that have a big impact and could prevent easy updating. We should wait for the next NuttX release. We want to update at some point for the UAV socket changes that were brought to NuttX from our community.
@dagar cleaned up mag drivers, some bosch driver was not using the calibration. He shifted them to use the PX4Magnetometer class. He improved the SPI setup to make it easier to shift drivers into the new structure. All IMU drivers can be shifted over except for I2C for now. If we find an easy way to include I2C we can get rid of the old drivers.
For now we want to test the pr that gets almost all boards into using the new drivers.
Magless GPS navigation (https://github.com/PX4/ecl/pull/770). Doc in https://docs.px4.io/master/en/advanced_config/tuning_the_ecl_ekf.html#yaw_from_gps_velocity
Currently you need to know where north is to fuse GPS which requires a calibrated compass. But with the new yaw estimator the north can also be infered from movement of the GPS position and this would allow flying without a magnetometer if the vehicles moves horizontally after takeoff. There are some limitations for now with having home, flying missions and having a quick position validity that still need to be fixed.
Pr enabling to fly Altitude without GPS was merged.
Acceleration setpoint execution / feed-forward is now merged: https://github.com/PX4/Firmware/pull/14212
Users that have customized configuration of
MPC_JERK_AUTO might want to lower them because of closer tracking. Users using the defaults for these parameters don’t require any changes.
Daniele found two bugs when repositioning with goto and change altitude: https://github.com/PX4/Firmware/pull/14543
MAVROS / RTPS / ROS2
Community contribution: Support for flightgear: https://github.com/PX4/Firmware/pull/14539
Support for Gazebo 11 is currently in work.
Roadmap, and Release discussion
@dagar We want to switch all boards over to the new drivers, make sure there are no regressions on that. And check that there are no remaining issues like PX4IO chip running out of RAM when loading larger airframe configurations like some VTOLs.
@JingerZ Table for classification of autopilots: https://px4.io/autopilots/
There is some grey zone and the table should make very clear which boards are directly supported, which ones are supported in cooperation with the manufacturers and which are experimental. @david_s5 Discussion about what means that the board supports the pixhawk standard, what if they support the fmu-v5 binary but have some small hardware differences for the peripherals?
@LorenzMeier Solve it by having clear categories:
- Pixhawk standard trademark: Electrically checked to work the same. If I plug things I’m sure it works on all these boards.
- PX4 fmu-v2-5 binary compatibility: Community/manufacturer supported. Mostly works like you would expect so it’s supported by PX4.
- Experimental: In the works, once introduced, untested.
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.