August 19, 2020
Meeting Time: 08:00 AM Pacific
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
Component update
Hardware Workgroups
Did not meet this week, there was the UAVCAN call.
Last updates: Hardware/Pixhawk Dev Call: Aug 11, 2020
System Architecture
No updates
OS / NuttX
@david_s5 is really close to having all the boards tested on the new NuttX version. Most things are contributed back upstream and made sure we preserve custom ports.
The qeustion is when to merge since the NuttX update should be brought in after the PX4 1.11 release. We should get and keep the new NuttX branch in a mergeable state and use it as much as possible for hardware tests.
@rroche has a question: The 1.11 release notes contain a NuttX upgrade but no further details. Answer: ~There was no update.~ There was an update @david_s5 will complete the changelog.
Driver
Commander
A lot of sensor calibration improvements open for testing, review:
Refactoring and logic improvbements
Estimation
No update
VTOL
No update
Fixed Wing
No update
Multicopter
Avoidance
Simulation
SITL testing on the sitl_gazebo submodule repository to catch problems early not only when they got merged in the submodule and should get updated in the Firmware repo:
Question: SITL tests in the Firmware seem to be failing quite a lot intermittently. We don’t know exactly where does problems are coming from. One hint of a problem is that activating offboard mode in the tests can have a race condition that’s also why iris fails, it’s the only model running offboard tests. We need to open an issue, collect the failure cases, track them down and not tolerate it breaking again.
MAVSDK
No updates
MAVROS / DDS / ROS2
MicroRTPS bridge update:
It’s now easier to use the bridge since you don’t have to rebuild it every time you want to use it.
UAVCAN
UAVCAN call updates: Hardware/Pixhawk Dev Call: Aug 18, 2020
Summary:
- DS-015 has better structure
- Physical layer specification/examples are in progress
- ESC message set is about to get finalized, please give your feedback now
- PX4 implementation: nunavut C serialization is currently blocking
If you are interested please join the call and give your feedback. All contributions appreciated.
Release
Release candidate 3 of PX4 v1.11 was release two days ago. The release is now branched off and no new major changes from master will automatically end up in v1.11. Stabilization of the release needs more testing but it’s close to be become the stable release. Bugfixes need to be cherry-picked when they get merged on master.
Please test by updating your vehicle from the last used version to v1.11 and report any regressions you notice in your usual use case.
Short highlight release notes need to be
Release blocker project board:
Community Q&A
Jacob Crabill: What’s the plan with all the different UAVCAN drivers? Right now we have a mix between between SocketCAN and the NuttX implementation. Our goal is to not carry an extra driver next to NuttX if there’s no important reason coming up. Peter gave an interesting presentation about the drivers in the NuttX workshop: https://www.youtube.com/watch?v=t49wq_ovT50&feature=youtu.be&t=19112
Follow up question about resource usage of the drivers? Any UAVCANv1 flavor will use significantly less memory than the current UAVCANv0 implementation.
Jim Donelson: Is a community member senior C/C++ dev experienced with ST MCUs looking into where to get started. He noted we have a problem with
Salim: Hi, I want to push forward this PR: https://github.com/PX4/Firmware/pull/15566
Or, does PX4 go into feature-freeze status and won’t accept new features until v1.11 is released?
Answer: Your change should be fine to be added looking at the diff.
Kalyan: Nora doesn’t have a PX4 IO chip. How does it work without AUX pins? With an octocopter that has additional servos? The support for all 14 PWM outputs was started but not finished yet. It should be resurrected and finished.
Is there a way to fix problems with running more than 4-6 DSHOT outputs? Usually they are DMA configuration limitations but on H7 it should possible to configure it the right way. An additional thing is using PWM outputs on the same rail which needs some work such that the right output driver gets started on the right hardware pins.
In-Depth discussions
-
https://github.com/PX4/Firmware/pull/15495
Possible short term solution: Just exclude the throttle and yaw and the main problem is gone.
Follow-up discussion: Have an RC and MAVLink stick input instance that handles gestures, handoffs.
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.