InstrumentStarCameraAngularAccelerometerFusion

This program estimates the satellites orientation from star camera data inputfileStarCamera and angular accelerometer data inputfileAngularAcc. The combination of both observation types is achieved in a least square adjustment. The optimal weighting between the two different observation groups is achieved by means of VCE in combination with a robust estimator. The system of linearized observation equations within the sensor fusion approach can be formulated as: \[ \begin{bmatrix} \M l_{ACC1B}\\ \M l_{SCA1B} \end{bmatrix} = \begin{bmatrix} \M A_{ACC1B} & \M B_{ACC1B}\\ \M A_{SCA1B} & \M 0 \end{bmatrix} \begin{bmatrix} \M q\\ \M b \end{bmatrix} = \begin{bmatrix} \frac{\partial \dot{\boldsymbol{\omega}}}{\partial \M q} & \frac{\partial \dot{\boldsymbol{\omega}}}{\partial \M b}\\ \M I & \M 0 \end{bmatrix} \begin{bmatrix} \M q\\ \M b \end{bmatrix} \]with \[\begin{split} \M l_{ACC1B} &= \dot{\boldsymbol{\omega}}_{ACC1B} - \dot{\boldsymbol{\omega}}_{0}, \\ \M l_{SCA1B} &= \M q_{SCA1B} - \M q_{0}, \\ \M q_{Fusion} &= \M q + \M q_{0}. \end{split}\]The reference values $\M q_{0}$ and $\dot{\boldsymbol{\omega}}_{0}$ are derived from inputfileStarCameraReference. In the course of the estimation, the accelerometer data is calibrated, by setting a bias factor $\M b$ with accBias.

Name	Type	Annotation
outputfileStarCamera	filename	combined quaternions
outputfileCovariance	filename	epoch-wise covariance matrix
outputfileCovarianceMatrix	filename	full arc-wise covariance matrix per arc. arc number is appended to filename
outputfileEpochSigmaStarCamera	filename	from vce and outlier detection
outputfileEpochSigmaAccelerometer	filename	from vce and outlier detection
outputfileAngularAcc	filename	angular acceleration observations (bias removed)
outputfileSolution	filename	estimated parameter (one column for each arc)
inputfileStarCameraReference	filename	quaternions as taylor point
inputfileStarCamera	filename	star camera observations
inputfileStarCameraCovariance	filename	star camera observations
inputfileAngularAcc	filename	angular acceleration observations
correctAccNonQuadratic	boolean	apply correction (non-square proof mass)
accBias	parametrizationTemporal	accelerometer bias per interval and axis
accScale	parametrizationTemporal	accelerometer scale per interval and axis
sigmaStarcamera	double	[rad]
sigmaAccelerometerX	double	[rad/s^2]
sigmaAccelerometerY	double	[rad/s^2]
sigmaAccelerometerZ	double	[rad/s^2]
estimateSigmaScaPerAxis	boolean	separate variance factor for roll, pitch, yaw, instead of one common factor.
estimateSigmaAccPerAxis	boolean	separate variance factor for each accelerometer axis, instead of one common factor.
huber	double	residuals > huber*sigma0 are downweighted
huberPower	double	residuals > huber: sigma=(e/huber)^power*sigma0
interpolationDegree	uint
iterationCount	uint	non linear equation solved iteratively

This program is parallelized.