DocumentCode :
1082291
Title :
On-line estimation of dynamic shock-error models based on the Kullback Leibler information measure
Author :
Krishnamurthy, Vikram
Author_Institution :
Cooperative Res. Center for Robust & Adaptive Syst., Australian Nat. Univ., Canberra, ACT
Volume :
39
Issue :
5
fYear :
1994
fDate :
5/1/1994 12:00:00 AM
Firstpage :
1129
Lastpage :
1135
Abstract :
Develops two sequential or “on-line” estimation schemes in the time domain for dynamic shock-error models which are special cases of errors-in-variables models. The author´s approach utilizes a state-space representation of the model, Kalman filtering techniques, and on-line algorithms. The first on-line algorithm is based on the expectation-maximization algorithm and uses a recursive Gauss-Newton scheme to maximize the Kullback Leibler information measure. The second on-line algorithm the author proposes is a gradient-based scheme and uses stochastic approximations to maximize the log likelihood. In comparison to the off-line maximum likelihood estimation scheme used in Ghosh (1989), the author´s on-line algorithms have significantly reduced computational costs and negligible memory requirements. Simulations illustrate the satisfactory performance of the algorithms in estimating errors-in-variables systems with parameters that vary slowly with time or undergo infrequent jump changes
Keywords :
Kalman filters; information theory; maximum likelihood estimation; numerical analysis; state-space methods; Kalman filtering techniques; Kullback Leibler information measure; computational costs; dynamic shock-error models; errors-in-variables models; expectation-maximization algorithm; gradient-based scheme; jump changes; log likelihood; negligible memory requirements; off-line maximum likelihood estimation; online estimation; recursive Gauss-Newton scheme; state-space representation; stochastic approximations; time domain; Computational efficiency; Computational modeling; Expectation-maximization algorithms; Filtering algorithms; Kalman filters; Least squares methods; Maximum likelihood estimation; Newton method; Recursive estimation; Stochastic processes;
fLanguage :
English
Journal_Title :
Automatic Control, IEEE Transactions on
Publisher :
ieee
ISSN :
0018-9286
Type :
jour
DOI :
10.1109/9.284909
Filename :
284909
Link To Document :
بازگشت