The quaternion to inverse rotation matrix function has been updated so that the rotation it produces is the inverse to that produced by the matrix library and the the inverse of the quaternion is uses. This function is now used to directly calculate an inverse rotation matrix rather than calculating the forward rotation and then transposing it.
- calculate a theoretical value based on the knowledge of the direction
and strength of the magnetic field vector and X/Y sensor measurements
- needs knowledge about location on earth to work
Fusion with large initial magnetometer biases errors can result in the the NE earth field states reducing in magnitude and effectively flipping sign.
EKF: Move declination state limiting into a separate function
EKF: Limit NE mag states after each 3-axis mag fusion
EKF: Fix bug in mag field strength look-up scale factor
Ensures that each time the earth field covariance and variance data is reset, that the off-diagonal elements containing earth field declination angle certainty is restored.
Moves calculation only required for mag heading fusion into the if (_control_status.flags.mag_hdg) branch
When using EV yaw, the observed yaw angle is calculated directly from the EV quaternions using derived expressions from references in code comments.
Everywhere where KHP is used, it is first completely reset, thus making
it unnecessary to keep it as a class member.
This saves 2.3KB RAM.
Stack sizes don't need changing, since there is already a function
Ekf::predictCovariance(), which needs around 3KB of stack and is called
close to where the fuse* functions are called.
Perform the innovation consistency check on all axes first and exit if any axis fails. Reduces the likelihood of bad magnetometer data corrupting the attitude estimates.
The code fragments for the magnetometer fusion have been refreshed from the auto-coder.
Explicit floating point types used for constants.
The 3x24 observation Jacobian has been replaced with a 1x24 that is updated each axis iteration to save memory.
The covariance was not being updated with the observation from one axis before the innovation variance was calculated for the next axis. This results in greater weighting on measurements for subsequent axes.
Paul Riseborough
RomanBapst
kritz
Carl Olsson
Daniel Agar
Julien Lecoeur
Beat Küng
Paul Riseborough