Now that we are using a consistent 50Hz minimum update rate for the covariance prediction we do not need a different initial gyro bias uncertainty for plane and copter to maintain filter stability margins.
The default value of 0.1 rad/s was too high and gave excessive settling time of the filter attitude after startup.
The initial attitude uncertainty has been increased to allow for some movement during startup.
If the baro data and magnetometer data are interleaved (arriving every 100 msec and offset by 50 msec), then the filter will go unstable during startup and fail to complete checks.
Fixes a potential error where changes to timing and arrival rate of magnetometer and baro data could block the fusion of synthetic position and velocity measurements, allowing unrestrained tilt errors during operation without GPS or optical flow.
Fusion of synthetic position or velocity measurements is now timed to coincide with fusion of barometer observations.
If a new barometer observation has not arrived after 200 msec then the synthetic position or velocity is fused anyway so that fusion of synthetic position or velocity observations cannot occur any slower than 5 Hz
Previous check default only checked the number of satellites and horizontal position accuracy.
Updated default value also checks HDoP and speed accuracy.
Now that we have a pre-arm check in place to detect bad lidar, the motion check is unnecessary and can false trigger for copters with flexible undercarriages or on uneven ground.
This method checks for consistency between accelerometer readings and switches to the unit with the lowest vibration of the difference exceeds 0.3g
The threshold of 1.7 m/s/s corresponds to a maximum tilt error of 10 deg assuming one IMU is good, one is bad and the EKF is using the bad IMU.
This commit changes the way libraries headers are included in source files:
- If the header is in the same directory the source belongs to, so the
notation '#include ""' is used with the path relative to the directory
containing the source.
- If the header is outside the directory containing the source, then we use
the notation '#include <>' with the path relative to libraries folder.
Some of the advantages of such approach:
- Only one search path for libraries headers.
- OSs like Windows may have a better lookup time.
If a badly conditioned covariance matrix causes negative innovation variances, then the filter will diverge. The previous approach of increasing process noise was not effective in some cases, so a hard reset of the covariance matrix has been adopted to guarantee recovery.
This fixes a numerical error observed using the replay on flight log which had significant periods of compass rejection.
Paul Riseborough
Andrew Tridgell
Gustavo Jose de Sousa
Randy Mackay
Jonathan Challinger
Tom Pittenger
Staroselskii Georgii
Paul Riseborough