Tradheli Autotune Improvements #26722
Conversation
|
@IamPete1 I made a pretty significant clean up of the heli autotune code. At least by my standards ;) I was wondering if you might take a look over the heli.cpp and provide any additional pointers on how to clean it up a little more. I'm sure there is a little more refactoring and ways that I could make it more readable. Already I have save 6K in flash on the cube orange+ :) |
|
@bnsgeyer you asked me to have a look at the autotests for this PR. It doesn't look like you're changing the test suite itself, just the simulated vehicle and default tuning values? I can have a look at the test failures, but if it's a vehicle dynamics sort of thing then you're far better off with help from @IamPete1 (or many other developers!) vs my rudimentary knowledge :-) |
|
@peterbarker sorrry, this was not the PR I was referring to. I just submitted the autotest PR last night. It is this PR I was having troubles with it but I think I figured out the error. Still would like you to look at it. |
bnsgeyer
force-pushed
the
pr-autotune-improve-tests
branch
from
b2d06ae to
0b3c2e1
Compare
bnsgeyer
force-pushed
the
pr-autotune-improve-tests
branch
5 times, most recently
from
fe46f83 to
cf14372
Compare
bnsgeyer
force-pushed
the
pr-autotune-improve-tests
branch
from
d8f8d38 to
ab0c00d
Compare
bnsgeyer
force-pushed
the
pr-autotune-improve-tests
branch
3 times, most recently
from
d37eb8f to
71183ec
Compare
| @@ -572,6 +607,7 @@ void AC_AutoTune_Heli::load_gain_set(AxisType s_axis, float rate_p, float rate_i | |||
| attitude_control->get_rate_roll_pid().slew_limit(smax); | |||
| attitude_control->get_angle_roll_p().kP(angle_p); | |||
| attitude_control->set_accel_roll_max_cdss(max_accel); | |||
| attitude_control->set_ang_vel_roll_max_degs(max_rate); | |||
There was a problem hiding this comment.
I don't think these are ever saved?
There was a problem hiding this comment.
These are only set to allow user to limit rates during autotune, if desired. autotune does not determine max rates.
There was a problem hiding this comment.
Do we need a load call to put it back to the user's value after the tune is done in that case?
There was a problem hiding this comment.
The load gain set method is used by the post test gain set method to load the original values. look above at the other methods that call the load gain set method. You have
load_tuned_gains
load_intra_test_gains
load_test_gains
load_orig_gains
in all cases except load_test_gains the orig rate max is used.
bnsgeyer
force-pushed
the
pr-autotune-improve-tests
branch
from
19605d2 to
6cf7b52
Compare
| void AC_AutoTune_FreqResp::set_dwell_cycles(uint8_t cycles) | ||
| { | ||
| dwell_cycles = cycles; | ||
| delete meas_peak_info_buffer; |
There was a problem hiding this comment.
Randy or PeterB will come up with a suggestion on how we can avoid the delete by create a larger-than-required ring buffer.
There was a problem hiding this comment.
@rmackay9 @peterbarker I removed the ring buffer delete statement and cleaned things up. I think my code works fine without them. I just clear the buffer before it is used the next time. I guess the buffer pops that first thing that was given to it if it is not full? Please have a look and let me konw.
Co-authored-by: Peter Hall <[email protected]>
bnsgeyer
force-pushed
the
pr-autotune-improve-tests
branch
from
00f6997 to
15e183b
Compare
|
This has been tested on actual helicopters by myself and @Ferruccio1984 |
This PR was primarily focused on enabling the safe tuning of larger helicopters. I have made other changes along the way that help clean up the code and make it more efficient. Here is a list of changes