If you take over 5 minutes for the question, please continue in Discord or a Github Issue.
Q.1: Crash report - Jw
Without companion computer, it flew perfectly. But after having companion computer (sending commands of 0?), it showed the weird behavior.
Q.2: ESC update through passthrough
This is a nice feature for sure.
BLHeli use case, we just need someone to implement it
Need to find contributor interested in this
Q.3: Reset speed when switching modes
Intuitively, it does make sense that we don’t *carry over the speed limit from another mode, to the new mode
But, we need more discussion on how to interpret this, what does it mean to change speed between flight modes?
Setting different cruise speeds for different modes of operation, is this what we desire?
E.g. we have “cruise speed” parameter for MC and FW, and for changing that (which should affect in long-term), it probably would be Service to change the parameter. But for temporarily changing the cruise speed shouldn’t incorporate parameter.
Geofence predictor is predicting the breach too early/wrong. Anyone with more context?
Setting ‘terimnate action’ can be very dangerous (e.g. lead to crash, as in the forum post), as there will be lots of false positive
This feature in general has a lot of false positives!
How can we improve this?
For now there’s quite a lot of Geofence issues
Q.4: Tilting multirotor support / testing
Implemented new Actuator Effectiveness Matrix method
How can we collaborate with PX4 team to get this incorporated properly?
In different labs, they rely on custom firmware, and has different problems. So using PX4 and fully supporting it would be much better for the research / industry!
@sfuhrer has enabled this in control allocation side, so could be nice to sync
This is for sure interesting. Cleanup to make it generic and not specific to this use-case would be better
Output of position is 3d force, and that gets interpolated by control allocator.
This would be a nice addition to control allocator
CMU had another PX4 fork that implemented 6 DOF implementation. We could check that out as well.
Can create a new branch and add as Pull Request in PX4-Autopilot main repository. That will show what diff exists & continue discussion- @salmarc
I’ve had a crash with my drone whilst flying in position mode with RC (without using the companion computer). I tried to switch back to manual but unfortunately could not save the drone anymore. Whilst analyzing the log files, I can’t directly find what might have caused the crash. I do see at the point where it went wrong a sudden jump in the roll and pitch set points, shortly after followed by one of the motors going to max thrust. The velocity controller seems to give some oscillations with a rather low damping. If there is anyone who might be able to shed some light on what might have gone wrong, it would be highly appreciated.
Also hoping to get a bit more information and context around Geofence Predictor in relation to my discussion post. This led to an early termination. It is a user error, but I can see others making easily making this mistake
how stable must the solution be to propose that to you? it has been tested but of course, there are things that will need to be checked or maybe changed
would it be possible to have someone collaborating on it in order to improve and perform other tests? I’m not a PX4 developer so of course there will be something wrong with it, but I’m trying to learn to improve
Is the RadioLink Ultrasonic Sensor SUI04 compatible with PX4 for obstacle sensing . The brochure indicates it is Mavlink / Pixhawk compatible OR do we require any special drivers here similar to TF LUNA setup from Benwake .
If you could start a forum post that would be great, but I think I heard you mention something about publishing certain topics from your onboard computer. Could you create a forum post and mention the message topics you are publishing from?