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

Copter: run rate loop at full filter rate in its own thread #27029

Open
wants to merge 19 commits into
base: master
Choose a base branch
from

Conversation

andyp1per
Copy link
Collaborator

@andyp1per andyp1per commented May 9, 2024

This is a redo of #26189

I have squashed the commits, rebased and started fixing the underlying problems. There were some fundamental problems with how the original PR was handling attitude control changes so I thought it was better to just open a new PR.

Support is enabled by setting:

FSTRATE_ENABLE = 1

which gives a variable attitude rate depending on load, and:

FSTRATE_ENABLE = 2

which gives an attitude rate locked to the gyro rate, but dynamic while disarmed. For testing there is also:

FSTRATE_ENABLE = 3

which is a locked rate at all times.

The output rate can be adjusted using FSTRATE_DIV which is a gyro divisor (default 1).
Two different rates - and therefore tunes - can be compared by using the lua applet switch_rates.lua which will persistently switch the appropriate tune/rate parameters with ones saved in names beginning with X_

ArduCopter/Attitude.cpp Outdated Show resolved Hide resolved
libraries/AC_PID/AC_PID.cpp Outdated Show resolved Hide resolved
@andyp1per andyp1per added the WIP label May 10, 2024
@andyp1per andyp1per force-pushed the pr-fast-rate-thread branch 6 times, most recently from e67b845 to 679611b Compare May 22, 2024 20:40
@andyp1per
Copy link
Collaborator Author

Flow today on top of 4.5 - working well.

@andyp1per andyp1per linked an issue May 26, 2024 that may be closed by this pull request
@andyp1per andyp1per force-pushed the pr-fast-rate-thread branch 13 times, most recently from 2370654 to 6398bd3 Compare June 7, 2024 19:29
@andyp1per
Copy link
Collaborator Author

andyp1per commented Jun 9, 2024

@rmackay9 this is close to being ready and I could use your input on what the configuration should look like. The basic premise is that we run the rate controller at the same rate as the gyros which is controlled by INS_GYRO_RATE (0=1Khz,1=2Khz,2=4Khz,3=8Khz). However, this is quite expensive and will not be possible on some hardware setups. In addition we need to be able to log the rate outputs at a faster rate than the main loop rate, but its unlikely that we could successfully log at the gyro rate. Finally we need to update the filters (notch filter in particular) at a faster rate than the main loop, but it does not make sense to do this any faster than at half the gyro rate.

Currently we have the following:
FLIGHT_OPTIONS=8 - enables the rate thread and dynamically schedules the relationship to the gyro rate based on load
FLIGHT_OPTIONS=16 - fixes the rate thread at the same rate as the gyro rate
ATT_FILTER_RATE - update rate of the filters in Hz or fixed at half the gyro rate if set to 0
ATT_LOG_RATE - update rate of the rate logs in Hz or set at the main loop rate if set to 0
The last two are not exact numbers - they are just the closest we can get to the required frequency by counting gyro samples so need to be an integer divisor or the gyro rate.

I'm not currently convinced about the naming or the configuration. I think this is an important enough feature that it needs its own enable flag. I also think it would be worth having a rate parameter that allows you to fix the rate. I also think that ATT is a little bit misleading. Maybe FAST_RATE_ or RATET_ or something. So an alternative might be:

FRATE_ENABLE - set to 1 to enable the rate thread
FRATE_RATE - set to 0 to dynamically throttle, fixed Hz otherwise
FRATE_LOG_RATE - as above
FRATE_FILT_RATE - as above

or something. It would also be possible to not use Hz at all but simply pick the divisor. So 2 would be gyro rate / 2 etc.
Your thoughts appreciated on this important topic!

ArduCopter/Attitude.cpp Outdated Show resolved Hide resolved
@andyp1per
Copy link
Collaborator Author

Has any testing been done without DShot? Does PWM still work I guess at it would still be at 400Hz. Do the DroneCAN ESC commands get sent faster with the higher update rate. Do the none motor outputs still work as expected (eg servo gimbal)?

I will need to do more testing. I believe this is all fine (including turtle mode), but will test further.

@andyp1per
Copy link
Collaborator Author

Turtle mode works fine

@andyp1per
Copy link
Collaborator Author

Verified that ESC commands are still going out over DroneCAN at 400Hz

Copy link
Contributor

@tridge tridge left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

please check we always pair cork/push together (@peterbarker suggested an internal error if we push without cork)

ArduCopter/motors.cpp Outdated Show resolved Hide resolved
@tridge
Copy link
Contributor

tridge commented Nov 13, 2024

i'm happy with the system level and the AP_InertialSensor and AP_Vehicle changes, so from the point of view of plane which doesn't use this I'm happy with this PR
I have not properly reviewed the copter specific changes

@tridge tridge removed the DevCallEU label Nov 13, 2024
andyp1per and others added 19 commits November 13, 2024 18:01
run motors output at rate thread loop rate
allow rate thread to be enabled/disabled at runtime for in-flight impact testing
setup the right PID notch sample rate when using the rate thread the PID notches
 run at a very different sample rate
call update_dynamic_notch_at_specified_rate() in rate thread
log RTDT messages to track rate loop performance
set dt each cycle of the rate loop thread
run rate controller on samples as soon as they are ready
detect overload conditions in both the rate loop and main loop
decimate the rate thread if the CPU appears overloaded
decimate the gyro window inside the IMU
add in gyro drift to attitude rate thread
add fixed-rate thread option
configure rate loop based on AP_INERTIALSENSOR_FAST_SAMPLE_WINDOW_ENABLED
better rate loop thread decimation management
ensure fix rate attitude is enabled on arming
add rate loop timing debug
update backend filters rather than all the backends
provide more options around attitude rates
only log attitude and IMU from main loop
force trigger_groups() and reduce attitude thread priority
migrate fast rate enablement to FSTRATE_ENABLE
remove rate thread logging configuration and choose sensible logging rates
conditionally compile rate thread pieces
allow fast rate decimation to be user throttled
if target rate changes immediately jump to target rate
recover quickly from rate changes
ensure fixed rate always prints the rate on arming and is always up to date
add support for fixed rate attitude that does not change when disarmed
only push to subsystems at main loop rate
add logging and motor timing debug
correctly round gyro decimation rates
set dshot rate when changing attitude rate
fallback to higher dshot rates at lower loop rates
re-factor rate loop rate updates
log rates in systemid mode
reset target modifiers at loop rate
don't compile in support on tradheli
move rate thread into its own compilation unit
add rate loop config abstraction that allows code to be elided on non-copter builds
dynamically enable/disable rate thread correctly
add design comment for the rate thread

Co-authored-by: Andrew Tridgell <[email protected]>
restrict rate loop to H7 and F7
honour the requested dshot rate as near as possible
…hread

decimate the gyro window locally
configure rate loop buffer based on AP_INERTIALSENSOR_FAST_SAMPLE_WINDOW_ENABLED
allow backends to be updated from rate thread
output debug error if rate loop buffer overruns
add support for updating filter parameters independently of propagating samples
add rate loop config abstraction that allows code to be elided on non-copter builds
must be using harmonic notch to use rate thread
mediate fast rate loop buffer using mutex and binary semaphore
ensure gyro samples are used when the rate loop buffer isn't

Co-Authored-By: Andrew Tridgell <[email protected]>
add nullptr checks and comments to FastRateBuffer
log after motor output in fast rate thread
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Run rate controller at full gyro rate
6 participants