Title :
The MIMO Discrete-Time Convergent Approximation of the Optimal Recursive Parameter Estimator
Author_Institution :
Electrical Engineering Department, University of California, Los Angeles, CA 90024-1594
Abstract :
A new discrete-time recursive parameter estimation algorithm is derived for a general multiple-input, multiple-output (MIMO) stochastic system that is bilinear in state and parameters. In contrast to the extended Kalman filter (EKF) for this system, the new algorithm can identify noise covariances and can be proven globally convergent. However, the new algorithm must also update approximations to third order moments, and so is more complicated than the EKF.
Keywords :
Approximation algorithms; Covariance matrix; MIMO; Nonlinear filters; Parameter estimation; Random sequences; Recursive estimation; Riccati equations; Stacking; Stochastic systems;
Conference_Titel :
American Control Conference, 1992
Conference_Location :
Chicago, IL, USA
Print_ISBN :
0-7803-0210-9
