Week 13 March 30 - April 3

Tony3dr · April 8, 2020, 2:56pm

Dronecode Test Team, Week 13 Report

PX4 Autopilot Flight Testing, by the Dronecode Test Team. Sponsored by Dronecode.

Summary

20 flights ([[[Link to log archive]
3 PR’s assigned
0 issue reported this week.

Vehicle Status

GitHub Pull Requests (Assigned / Tested)

magnetometer’s move to PX4Magnetometer and cleanup [#14536]
Enable GPS flight without magnetometer #14162
ekf2 handle accelerometer clipping #14304

github.com/PX4/PX4-Autopilot

ekf2 handle accelerometer clipping

PX4:master ← PX4:pr-ekf2_accel_clipping

opened 05:17PM - 05 Mar 20 UTC

dagar

+101 -36

- pass accel clip counts through sensors hub to `sensor_combined` - copy accel… clip count (and total samples) into ecl/EKF IMU buffer - ecl/EKF increase delta velocity variance while IMU sample contains any clipping (https://github.com/PX4/ecl/pull/663) - tunable via new parameter `EKF2_ACC_CLPFACT` Requires https://github.com/PX4/ecl/pull/663 Example flight log with clipping. https://github.com/PX4/Firmware/pull/14294#issuecomment-594918621 ![clipping](https://user-images.githubusercontent.com/84712/76006835-161bf080-5edb-11ea-88e8-d9761087d337.png) TODO: - [ ] get representative logs with clipping (test team pixracer and/or pixhawk 3 pro vehicles) - [ ] review and tune via ekf2 replay The goal is make these brief periods of accel clipping survivable, including alerting the operator if necessary rather than getting caught in the death spiral of clipping -> vz jumps -> more throttle -> more clipping, etc.

Week 13 March 30 - April 3

Dronecode Test Team, Week 13 Report

Summary

Vehicle Status

GitHub Pull Requests (Assigned / Tested)

Issues reported: 0