Problem areas and practical solutions in the development of large-dimension Kalman filters for the calibration and alignment of complex inertial guidance systems are discussed. The basic vector attitude error equation is augmented by gyro and accelerometer unknown parameters. The parameter estimation problem is converted into a state estimation problem. A complete approach and description of the dual extended Kalman filter, one for accelerometers and one for gyros, is given. To reduce computational load, a technique of prefiltering (data compression or measurement averaging) has been implemented in the mechanization with very little degradation in the performance of the filter. The models of gyros and accelerometers used are described in detail. A technique for generating parameter excitation trajectories which provides observability of instrument parameters has been developed by maximizing the information matrix. A typical set of results for a simulator data set for parameter estimates and innovation sequences is given to show the performance, convergence, accuracy, and stability of the filter estimates.<<ETX>>
[1]
R. Mehra.
Optimal inputs for linear system identification
,
1974
.
[2]
David M. Auslander,et al.
Control and dynamic systems
,
1970
.
[3]
Cornelius T. Leondes,et al.
Control and Dynamic Systems
,
1985
.
[4]
M. Grewal,et al.
Optimal selection of trajectories for parameter estimation in dynamical systems
,
1982,
1982 21st IEEE Conference on Decision and Control.
[5]
G. A. Napjus,et al.
Design of optimal probing signals for vector parameter estimation
,
1971,
CDC 1971.
[6]
Raman K. Mehra,et al.
Optimal input signals for parameter estimation in dynamic systems--Survey and new results
,
1974
.
[7]
G. H. Myers.
Guidance and control of aerospace vehicles
,
1965
.
[8]
K. Glover,et al.
Relationships between identifiability and input selection
,
1975,
1975 IEEE Conference on Decision and Control including the 14th Symposium on Adaptive Processes.