My four meter craft is achieving the desired turn rate well, but the so called ‘desired’ turn rate is far from desirable.
Surely a truly desired turn rate should follow a straight line from waypoint to waypoint, not a drunken like wobble from wpt to wpt.
What parameters will bring the desired turn rate closer to the actual mission (straight line wpt to wpt)?
LOG - https://drive.google.com/file/d/1b4N3t4kdnz6QAyui3Ux5o0fPBNhYcGbu/view?usp=sharing
Not a turn rate problem, rather a position control problem (the PSC_ parameters). Without a log, impossible to help further.
I’ll have a look, but I make no promises. Tuning the position controller has proven quite tricky (and is actually holding up the stable release of Rover 4.3).
I look forward to @rmackay9’s analysis if he has a bit of time to provide it.
Zooming to a section of the log where I think a straight path is being attempted shows a lot of position controller overshoots.
P is oscillating quite a bit while the I term is not saturating, and the D term is contributing very little (despite the noise), so I think reducing P is the ticket here. Maybe reduce PSC_VEL_P from 1 to 0.5 and see how that goes?
The rate tuning is somewhat well behaved, though there is quite a bit of noise in the speed controller. Reducing ATC_SPEED_P and I may help a bit with that.
Thank you for that Yuri. Do you perhaps have a link to detailed documentation explaining the tuning of the position controller for Rover 4.3? I really need to get my head around this.
Have a look at this section of the log where the throttle is almost shut right down for absolutely no reason at all. This throttle issue has been a hurdle from day one with Rover 4.3.
I’m not sure I can fully explain that behavior, but I suspect it’s related to the poor navigation performance/position controller tune.
Here is the documentation for tuning 4.3 nav:
Tuning Navigation — Rover documentation (ardupilot.org)
Thank you Yuri, I hadn’t noticed that it had been added.