Title :
EM algorithm convergence for inertial navigation system alignment
Author :
Einicke, Garry A.
Author_Institution :
Commonwealth Sci. & Ind. Res. Organ. (CSIRO), Pullenvale, QLD, Australia
Abstract :
The convergence of a Kalman filter-based EM algorithm for estimating variances is investigated. It is established that if the variance estimates and the error covariances are initialized appropriately, the sequence of variance iterates will be monotonically nonincreasing. Under prescribed conditions, the variance estimates will converge to the actual values. An inertial navigation application is discussed in which performance depends on accurately estimating the process variances.
Keywords :
Kalman filters; convergence of numerical methods; covariance analysis; error analysis; expectation-maximisation algorithm; inertial navigation; Kalman filter-based EM algorithm convergence; error covariance; expectation-maximisation algorithm; inertial navigation system alignment; variance estimation; variance iterate sequence; Control systems; Convergence; Difference equations; Inertial navigation; Iterative algorithms; Maximum likelihood estimation; Noise measurement; Parameter estimation; Riccati equations; State estimation; Kalman filtering; inertial navigation; parameter estimation; stationary alignment;
Conference_Titel :
Decision and Control, 2008. CDC 2008. 47th IEEE Conference on
Conference_Location :
Cancun
Print_ISBN :
978-1-4244-3123-6
Electronic_ISBN :
0191-2216
DOI :
10.1109/CDC.2008.4739113
