farhang
September 24, 2024, 4:57am
1
PX4 Sync / Q&A: Sep 25, 2024
Announcements
Release Discussion
Q&A
We have a new release :
PX4 Autopilot, the open-source flight control system for uncrewed vehicles, has released its latest stable version, v1.15. This update introduces significant improvements for PX4 developers, including tighter integration with ROS 2 and uXRCE-DDS...
Est. reading time: 6 minutes
New documentation section:@hamishwillee and @Vincentpoont2
GPS Issue fixed and back-ported:
PX4:release/1.15 ← PX4:fix-septentrio-default-1_15
opened 09:23AM - 25 Sep 24 UTC
backport of https://github.com/PX4/PX4-Autopilot/pull/23718
fixes https://githu… b.com/PX4/PX4-Autopilot/issues/23714
1- Firmware v1.15 stable _ GPS not detected _ CUBE orange _ Herelink
2- @anaam-wingxpand
opened 01:36PM - 25 Sep 24 UTC
bug-report
### Describe the bug
Landing was detected but did not disarm, despite COM_DISAR… M_LAND = 2 seconds. In fact, 5 seconds later the motor started again as takeoff was detected, even though it wasn't actually taking off.
### To Reproduce
1. The operator shook the drone to pre-trigger launch.
2. The operator held onto the drone. 2 seconds later, the motor stopped as landing was detected
3. The operator went to go set the drone down, and 5 seconds later the motors started again as takeoff was detected
4. The operator disarmed the drone via QGroundControl
### Expected behavior
The drone should've disarmed 2 seconds after landing, since COM_DISARM_LAND = 2 seconds. The motors should never started up again on their own.
### Screenshot / Media
### Flight Log
https://logs.px4.io/plot_app?log=924fa87f-0220-43ae-b9c5-b21d753078f5
### Software Version
v1.14
### Flight controller
Cube Orange+
### Vehicle type
Fixed Wing
### How are the different components wired up (including port information)
_No response_
### Additional context
_No response_
3- Is there any roadmap for the future changes at architecture level for upcoming PX4 releases?
4- @Rowan_Dempster Huge drift on GPS after first time flying an indoor drone, outdoors with vision sources for localization.
Merged:
PX4:main ← AlexKlimaj:pr-fix-corrupt-log-parsing
opened 02:24AM - 24 Sep 24 UTC
Currently any log with a minor corruption will not upload and complain about the… log being corrupt. Plotjuggler opens the logs without issue. Allowing string exceptions fixes this.
```
./serve.py --show
/home/alex/.local/lib/python3.10/site-packages/matplotlib/projections/__init__.py:63: UserWarning: Unable to import Axes3D. This may be due to multiple versions of Matplotlib being installed (e.g. as a system package and as a pip package). As a result, the 3D projection is not available.
warnings.warn("Unable to import Axes3D. This may be due to multiple versions of "
Moving uploaded file to /home/alex/flight_review/app/plot_app/../../data/log_files/10c53e34-3767-493b-ba84-20332f3d225f.ulg
Traceback (most recent call last):
File "/home/alex/flight_review/app/plot_app/helper.py", line 317, in load_ulog_file
ulog = ULog(file_name, msg_filter, disable_str_exceptions=False)
File "/home/alex/.local/lib/python3.10/site-packages/pyulog/core.py", line 137, in __init__
self._load_file(log_file, message_name_filter_list, parse_header_only)
File "/home/alex/.local/lib/python3.10/site-packages/pyulog/core.py", line 836, in _load_file
self._read_file_definitions()
File "/home/alex/.local/lib/python3.10/site-packages/pyulog/core.py", line 886, in _read_file_definitions
msg_info = self._MessageInfo(data, header, is_info_multiple=True)
File "/home/alex/.local/lib/python3.10/site-packages/pyulog/core.py", line 563, in __init__
self.value = ULog.parse_string(data[1+key_len:])
File "/home/alex/.local/lib/python3.10/site-packages/pyulog/core.py", line 91, in parse_string
ret = _parse_string(cstr)
File "/home/alex/.local/lib/python3.10/site-packages/pyulog/core.py", line 17, in _parse_string
return str(cstr, 'utf-8', errors)
UnicodeDecodeError: 'utf-8' codec can't decode byte 0xba in position 152: invalid start byte
400 POST /upload (127.0.0.1): Failed to parse the file. It is most likely corrupt.
400 POST /upload (127.0.0.1) 153.78ms
```
We would do this on a clean system with CI:
PX4:main ← PX4:pr-macos-build-rev1
opened 03:32PM - 12 Sep 24 UTC
This is initial sketch. Let me know if I am missing anything here!
PX4:main ← dakejahl:pr-serial_drain_tx
opened 03:02AM - 18 Sep 24 UTC
Calling serial::write() in quick succession was blowing away the previous buffer… , fsync does not guarantee that data is transmitted on serial lines. On the other hand tcdrain waits until the output buffer is empty.
I tested on nuttx, probably need to update posix as well. @katzfey was the intention with fsync to flush the output buffer? I discovered this issue in a custom module where I am updating the firmware of a device, reading a firmware file from the file system and writing chunks of 512 bytes at a time. I was discovering that it was rapidly looping and not actually writing each 512 byte chunk to completion.
PX4:main ← cuav:cuav_7-nano
opened 08:06AM - 16 Jul 24 UTC
Added support for CUAV-7-Nano flight controller.
Now that PX4 v1.15 stable is out, maybe there’s a window to plan and make bigger changes/updates.
My question is the following: are there any plans related to upgrading NuttX anytime in the near future?
No pressure, I was just curious, I haven’t found any information, just one fairly old GitHub issue:
https://github.com/PX4/PX4-Autopilot/issues/22269
@dagar : Are there any news related to this topic?
