delete control_state and cleanup vehicle_attitude #7882
Conversation
| @@ -166,6 +168,7 @@ class MulticopterAttitudeControl | |||
| struct battery_status_s _battery_status; /**< battery status */ | |||
| struct sensor_gyro_s _sensor_gyro; /**< gyro data before thermal correctons and ekf bias estimates are applied */ | |||
| struct sensor_correction_s _sensor_correction; /**< sensor thermal corrections */ | |||
| struct sensor_corrected_s _sensor_corrected; /**< sensor in-run bias corrections */ | |||
There was a problem hiding this comment.
@dagar sensor_correction and sensor_corrected is a terrible choice - could we change to sensor_bias instead? Because that is what it really is.
There was a problem hiding this comment.
That came in after what I originally did, but it sounds fine to me. There are a few users of the bias correct accel though.
I understand why we're doing it, but this entire sensor reconstruction in the rate controller feels awkward. I'll do the bias rename for now, but I'd like to discuss other options later on.
|
@PX4TestFlights Please re-test. Thanks! |
|
couple flights with pixracer (v4) good flights; steady flights; no notable difference from current master |
|
flight with pixhawk 1 (V2) flight with pixhawk mini (V3) good flights, no issues |
|
today flight with pix (v4-pro) |
|
Couple flights with pixhawk pro (V4) VTOL Good flights no issues! Couple flights with pixhawk mini (V3) FIXED WING Good flights no issues! |
dagar
force-pushed
the
pr-att-bias-cleanup3
branch
from
2988c23
to
8c2411e
Compare
|
@dagar i will try to fly this tomorrow, can you rebase? |
|
Done |
dagar
force-pushed
the
pr-att-bias-cleanup3
branch
from
8c2411e
to
7999881
Compare
|
Rebased yet again. Did you get a chance to test this @sanderux? |
|
Sorry, no, while in the field we had all but a usb cable. in the next few days i will. |
dagar
force-pushed
the
pr-att-bias-cleanup3
branch
from
7999881
to
a3f0e19
Compare
|
Rebased again. |
dagar
force-pushed
the
pr-att-bias-cleanup3
branch
from
a3f0e19
to
c6ba062
Compare
|
@dagar Do you need @PX4TestFlights to test this again? |
|
I think we're mostly good on testing other than VTOL without an airspeed sensor. |
|
@dagar would you mind rebasing again? |
dagar
force-pushed
the
pr-att-bias-cleanup3
branch
from
c6ba062
to
ca62230
Compare
There was a problem hiding this comment.
No regression on airspeedsensorless
https://logs.px4.io/plot_app?log=0200995e-1ff9-4d86-8d94-c19ca288c88c
|
Finally! Woohoo! |
|
@dagar could you please summarize how this impacts sensor drivers? The diff is pretty big and is all in one commit. |
|
@nicolaerosia this was mainly a topic cleanup. There was a significant amount of duplicated data being published by the estimators in a topic called control_state. I apologize for the enormous diff. This dragged out for a long time across multiple people and rebasing became unmanageable. It shouldn't have any impact on sensor drivers. |
Continuation of #7773. The rebase was becoming too difficult. I can split this if necessary.