Conversation
- Use percent error for comparing RMS error values
|
Looks like a good start. Not that we've defined this anywhere, but would you mind running it through PX4's astyle? |
|
These changes appear to be causing a problem with PX4/Firmware CI testing. |
|
It looks like they are passing in this CI: |
|
My initial thought is that it appears the |
|
I'll limit the ecl tests to the standalone build. #591 |
Bug Fixes: PX4/PX4-ECL#586 - EKF: fix covariance and output filter buffer initialization PX4/PX4-ECL#590 - EKF: Fix innovation in fuseDeclination() Enhancements: PX4/PX4-ECL#543 - ekf_helper: add more useful methods to interface with the covariances PX4/PX4-ECL#588 - Add unit tests for DataValidator Note: PX4/PX4-ECL#543 has required a change to how the state variances are accessed .
Bug Fixes: PX4/PX4-ECL#586 - EKF: fix covariance and output filter buffer initialization PX4/PX4-ECL#590 - EKF: Fix innovation in fuseDeclination() Enhancements: PX4/PX4-ECL#543 - ekf_helper: add more useful methods to interface with the covariances PX4/PX4-ECL#588 - Add unit tests for DataValidator Note: PX4/PX4-ECL#543 has required a change to how the state variances are accessed .
Added a few unit tests for the
DataValidatorclass that cover happy path and common failure paths. In some cases the existing interface lacks documentation of expected behavior (for exampleDataValidator::confidencehas some undocumented side-effects, setting_error_mask, andputperforms a variety of calculations), so these tests codify some of the "as-built" behavior of the class. I did not yet attempt to test the vibration-calculating code, which is unspecified but appears to be key to the operation ofDataValidatorGroup.