Fix flow gyro bias corrections #22145
Conversation
|
This pull request has been mentioned on Discussion Forum for PX4, Pixhawk, QGroundControl, MAVSDK, MAVLink. There might be relevant details there: https://discuss.px4.io/t/optical-flow-oscillating-position/34470/2 |
Prevents from applying bias corrections several times if the sample isn't immediately used.
bresch
force-pushed
the
pr-flow-fix-gyro-compensation
branch
from
42df747
to
d002826
Compare
bresch
changed the base branch from
pr-ekf2_optical_flow_log_gyro_bias
to
main
| @@ -79,7 +79,7 @@ void UavcanFlowBridge::flow_sub_cb(const uavcan::ReceivedDataStructure<com::hex: | |||
| if (PX4_ISFINITE(msg.rate_gyro_integral[0]) && PX4_ISFINITE(msg.rate_gyro_integral[1])) { | |||
| flow.delta_angle[0] = msg.rate_gyro_integral[0]; | |||
| flow.delta_angle[1] = msg.rate_gyro_integral[1]; | |||
| flow.delta_angle[2] = 0.f; | |||
| flow.delta_angle[2] = NAN; | |||
There was a problem hiding this comment.
Hmm, with this it's still going to replace the entire gyro as it passes through sensors/vehicle_optical_flow.
There was a problem hiding this comment.
Good point. Should we rather do the gyro Z stitching in VehicleOpticalFlow directly and remove it from EKF2?
There was a problem hiding this comment.
I'm not sure what's best overall, but at the moment one reason not to do it in sensors/vehicle_optical_flow is because we're not currently accounting for any optical flow delay there.
|
Flight review is down. But here are two logs. One with GPS and one without. Overall pretty good. Did something a bit funny on the optical flow only flight when I was reversing straight back it veered off course aggressively. https://review.px4.io/plot_app?log=6c83b40c-a680-475b-b26b-b2ac2f9fcab6 |
|
This pull request has been mentioned on Discussion Forum for PX4, Pixhawk, QGroundControl, MAVSDK, MAVLink. There might be relevant details there: https://discuss.px4.io/t/optical-flow-oscillating-position/34470/5 |
Add the autopilot's Z gyro when this is the case
bresch
force-pushed
the
pr-flow-fix-gyro-compensation
branch
from
d002826
to
bf0b369
Compare
|
Bench test showing how the Z angular rate is added to the flow gyro data (in center plot, the Z gyro is NAN) |
|
This pull request has been mentioned on Discussion Forum for PX4, Pixhawk, QGroundControl, MAVSDK, MAVLink. There might be relevant details there: https://discuss.px4.io/t/holybro-optical-flow-sensor-thoneflow-oscillations-on-main/34768/1 |
* ekf2-flow: fix flow gyro bias compensation * ekf2-flow: apply flow gyro bias when used * ekf2: log optical flow gyro bias * ekf2: optical flow control always use provided flow gyro (with bias applied) * ekf2-flow: log flow gyro and gyro reference * ekf2-flow: support senrors with XY flow gyro --------- Co-authored-by: Daniel Agar <daniel@agar.ca> Co-authored-by: bresch <[brescianimathieu@gmail.com](mailto:brescianimathieu@gmail.com)>
Solved Problem
Position oscillations with ArkFlow
Solution
Changelog Entry
For release notes: