April 08, 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
There was an issue when you unplugged USB but the board didn’t reboot it would busy loop.
If the board is not armed and USB is unplugged it would reboot since https://github.com/PX4/Firmware/pull/3053 to free up any resources implicitly. It’s not an actively supported use case to change that.
Testing ARM GCC 9 builds, we’ll not move to the new compiler for the v1.11 release:
OS / NuttX
- F7 SPI DMA merged and backported.
- FlexCAN merged, needs testing on further boards
- All boards are now on the new Invensense drivers
It’s using the unified driver architecture, supports changing the update rate.
- Proposal to lower the IMU gyroscope update rate because CPU load on certain boards is high with 1kHz update rate. We should see if we can manually correctly set the parameters on the affected boards making sure the rest of them can profit from the high update rate.
- Fix for setting the home position: Previously the home position was reset when arming, now it will reset when you take off.
- Main thin contribution from @dagar when the driver detects clipping we increase the variance of the sensor such that it doesn’t influence the estimate strong anymore.
- @Paul_Riseborough found some numerical issues with the accelerometer bias estimation but it’s not a regression and really hard to trigger, requires hours of standing in an unobservable state.
@maetugr Possible regression with acceleration feed-forward after transition.
New gimbal protocol implementaiton, work in progress contribution by Kalyan Sriram from AmadorUAVs:
Language wrappers were auto generated but C++ was written manually. Result C++ had more functionality. Now C++ can be auto generated as well but and there’ll be feature parity for everything auto-generated.
MAVROS / RTPS / ROS2
The way models were spawned in SITL was changed and caused problems in SITL tests in CI:
There was a race condition in the gimbal control communication:
Roadmap, and Release discussion
Component maintainers, check your components and make sure it’s stable and be aware of blockers: https://github.com/PX4/Firmware/projects/17
RAM situation on IO? Some rare configurations can still fail to load. @dagar has an idea to save some RAM such that we can get through the release with the mixers as they are.
@nicovanduijn Defaults change proposal with not a lot of feedback yet: https://github.com/PX4/Firmware/pull/14560
It was tested on a large airframe and the PX4 vision. It’s usually nice to have a log posted.
Going through release board in detail: https://github.com/PX4/Firmware/projects/17
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.