Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

PX4 V1.9.0 and Later: Attitude Control and Takeoff Issues #12484

Closed
ar1040 opened this issue Jul 15, 2019 · 13 comments
Closed

PX4 V1.9.0 and Later: Attitude Control and Takeoff Issues #12484

ar1040 opened this issue Jul 15, 2019 · 13 comments

Comments

@ar1040
Copy link

ar1040 commented Jul 15, 2019

@mcsauder as you requested. I have a fleet of quads that have been flying extremely well for months now. I am using a Pixhawk Cube with a Here+ GPS. I updated one of these units to v1.9.0, and this is where I found this problem. I have also flashed v1.9.1 and v1.9.2 all with the same problems:

Takeoff: when the vehicle lifts off it seems to do so very weakly, as opposed to taking off quickly. About 2 out of 3 times, the vehicle looses heading (leaving the ground) and corrects itself during ascent, or as it reaches the takeoff height. The issue for me, is my quads often carry heavy flat payloads in high wind situations and I am afraid this may cause a crash.

Attitude control: The vehicle is controlled via companion computer relaying to a ground station allowing easy keyboard controls. We had our craft slowed down to restrict jerky movement and to restrict yawing speeds (as I said heavy flat payloads). When I updated to v1.9.0 and up, the yawing became extremely fast (almost catastrophic). Also jerking the vehicle side to side was also very stable prior to updating, but with the newest update version, the drone almost flips itself when jerking side to side (without a payload).

I have tried this on 4 different units all with the same issue.

I would prefer to keep my vehicles up to date, but I have been sticking with v1.8.2 until the issues have been resolved.

@julianoes
Copy link
Contributor

@MaEtUgR @bresch is this a regression with smooth takeoff?

@mcsauder
Copy link
Contributor

See also #12278

@ar1040
Copy link
Author

ar1040 commented Jul 16, 2019

If any other resources (ie. logs, videos, etc) are needed let me know and I will record and upload.

@mcsauder
Copy link
Contributor

If you can record a flight log on your system with 1.8 and another with 1.9.2 it would be helpful. I will do the same on my system and we can look through them as an A/B comparison.

@ar1040
Copy link
Author

ar1040 commented Jul 17, 2019

Of course when I reach out to the Devs my craft wants to takeoff fine. I had no yawing on takeoff on the first drone. I will try a couple more today in order to replicate the problem. However, on my last takeoff, the drone disarmed itself before/at the takeoff height. I thought I would mention this as it still has to do with takeoff. Here is the log for that crash:

https://logs.px4.io/plot_app?log=2785e143-48b4-4b78-b85c-692afe2a79a1

@mcsauder
Copy link
Contributor

@ar1040 , from your log above it appears that at the transition from offboard control back the autopilot thought the vehicle was landed. I'm not familiar enough with offboard control to know what could have happened in that transition.
image

If you can, post a log with the attitude issue when you encounter it again and if you can get a video recording that couldn't hurt to look at either. I cannot reproduce the issues you have described with 1.9.2 or current master, so I'm in the dark about what could be the cause.

@julianoes julianoes added the bug label Jul 18, 2019
@ar1040
Copy link
Author

ar1040 commented Jul 18, 2019

@mcsauder , I did have a chance to test the attitude on that drone before it crashed.

https://logs.px4.io/plot_app?log=63e9657f-7457-4fba-b6c8-d3ce86af8683
https://logs.px4.io/plot_app?log=326d4b5c-4e40-4e69-a5e0-cc342e1d0bd9
https://logs.px4.io/plot_app?log=9a29f67a-955c-402f-93f3-cc33f577e416

I gave the craft multiple maneuvers while testing. There are some hard side-to-side maneuvers included. During these hard side-to-side maneuvers, the drone acted like I described earlier, acting as it may flip, and becoming overly responsive. Looking through the logs, I couldn't see much of a difference in the roll angle like I expected, but maybe you could see more than I can.

Edit: On the subject of the disarming, I would like to add that this hasn't happened to me using version 1.8.2. I was wondering is there was a transition state in the firmware, so when the drone is prompted with a takeoff command, basically so it cannot disarm itself while in the air. Looking through our code, we have explicit checks that the drone may not disarm itself other than when in a landed state, and the only way to reach a landed state is by giving the drone a land command, which cannot be executed until 5 seconds after the craft reaches its takeoff altitude.

Edit2: We do however have a check where, when the drone takes off, it has to reach a specific height in 5 seconds, or the drone is told to land. After being told to land it tries to disarm itself, but its normally not able to disarm itself before it has landed.

@ar1040
Copy link
Author

ar1040 commented Jul 18, 2019

I also wanted to add, I mistakenly had a drone with the version 1.9.0, and the same thing happened, it fell from the sky during takeoff. This has not been a problem for me when using 1.8.2, and I am not sure of the differences. Here's the log to that crash to see if there is any consistency:

https://logs.px4.io/plot_app?log=0ac3974c-0501-44b9-98b4-f70a4ec885c7

@stale
Copy link

stale bot commented Oct 16, 2019

This issue has been automatically marked as stale because it has not had recent activity. Thank you for your contributions.

@stale stale bot added the stale label Oct 16, 2019
@tonight123
Copy link

I have the same hardware, firmware, as you have, and I have the same problem。
takeoff: mpc_tko_ramp_t =0,it's going to take off quickly.
Starting counterclockwise after take-off, probably because rate i is too big, i is currently 0.2

@stale
Copy link

stale bot commented Jan 21, 2020

This issue has been automatically marked as stale because it has not had recent activity. Thank you for your contributions.

@stale stale bot added the stale label Jan 21, 2020
@julianoes
Copy link
Contributor

@ar1040 were you able to get to the bottom of this or fix it?

@stale stale bot removed the stale label Jan 24, 2020
@ar1040
Copy link
Author

ar1040 commented Jan 24, 2020

Unfortunately, we had to mode on. We stuck to using the latest commit of v1.8.2. This version seems to have minimal issues. I will be looking forward to testing v1.10 in the near future, and I will reopen this if the problem persists. Thank you all for your help.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

4 participants