This operation is expensive when done to the whole covariance matrix and unnecessary after covariance prediction because we calculate the upper diagonal and copy across so it is already symmetric.
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.
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
Simplify calling so that it is only called in two ways:
1) Immediately before 3-axis mag fusion if not called since the last earth field covariance reset so that the earth field declination information can be formed.
2) Immediately after 3-axis mag fusion otherwise.
In the case where the EKF is switching between 3-axis and heading fusion, off-diagonal elements containing the correlation between N,E components of the earth field were being lost on each switch event. These elements contained information about the declination uncertainty and should be preserved.
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.
Use a new method that preserves the roll and pitch information and adds the uncertainty for yaw only.
Ensure that correlation information to non-quaternion states is removed when a reset occurs to prevent fusion of subsequent observations (e.g. GPS) causing incorrect yaw.
The handling of invalid flow data when on ground is performed in controlOpticalFlowFusion() where it is able to handle flow sensors that don't publish gyro data.
Heading data is assumed to be from a dual antenna array at a specified yaw angle offset in body frame, but with the heading data already corrected for antenna offset. The offset is required to apply the correct compensation for combined rotations and to determine when the yaw observation has become badly conditioned.
The parameter used to control the maximum dead reckoning time had 'gps' in the parameter name which was confusing because it was used for all measurement types capable of constraining horizontal velocity error growth. The parameter variable has been renamed and the documentation for it improved.
The parameter used to control the maximum time since fusing a measurement before the measurement is considered to be not contributing to aiding had misleading documentation which has been updated.
The high frequency acceleration noise levels in the gazebo models and also seen on some hardware is causing the IMU vibration check warning to fail.
The thresholds have been lifted and the reporting improved to make it clearer which sensor noise is causing the failure.
Motion compensated optical flow rates are supposed to be zeroed if reported flow quality is below the minimum threshold value when on ground.
The comments and logic have been amended to be consistent and make the design intent clearer.
Terrain validity is determined solely by successful range finder fusion and terrain state initialisation.
A range finder that has been declared faulty requires continuous range finder data fusion requires data to be continuous before the fault status _rng_hgt_faulty can be cleared. This will enforce the requirement for continuous data before fusion can commence.
Eliminate race condition caused by checking for data freshness using time stamps from buffer push instead than buffer pop events.
Consistent use of range data ready and range data fault flags. This achieved by ensuring that _rng_hgt_faulty is set to true for all range data faults, not just data freshness.
Include range data validity requirement in rangeAidConditionsMet() check.
Squashed commits:
[ed2a243] FlightTasks: Preserve control loop tuning when applying max altitude limit
[b33b947] FlightTasks: Add terrain hold function
This new mode of altitude control uses terrain following when holding position and normal altitude control when moving.
This brings all the range finder data checks (excluding innovation consistency checks) into one place and eliminates the need to perform range checking external to the library.
The hard coded optical flow tilt limit is changed to use the same value as the range finder.
Variable names are changed to make a clear distinction between the max/min values calculated by the stuck range check and the max/min valid values for the sensor.
Uses a relaxed yaw innovation check threshold for fixed wing vehicle that can recover from larger yaw errors after takeoff.
Replaces the peak hold with decay filtering applied to the magnetic yaw innovations with a conventional lowpass filter. This prevents mag heading innovation transients caused by preflight handling from failing the check, but catches persistent errors. Sign swapping of innovations due to angle wrapping is not a problem due to the way that innovations are calculated inside the EKF, so a simple LPF sufficient.
The gyro data accumulation needs to be across the same integration period as the flow sensor. The previous code didn't sample the accumulation until the midpoint of the flow data had fallen behind the fusion time horizon.
This PR changes the optical flow time stamp definition so that flow data is retrieved when the leading edge of the flow accumulation period falls behind the fusion time horizon. This enables the accumulated gyro data to be sampled at the correct time. Fusion is then delayed until the mid sample time has fallen behind the fusion time horizon.
High rate optical flow data could make flow fusion to run every major update cycle, resulting in the calculation of bias errors in the body rates used to compensate flow data failing time validity checks and not running. This resulted in a slow drift of the nav solution if bias errors were present in the in the gyro data used for flow sensor motion compensation.
Paul Riseborough