Numerical experiments with various optimal estimators |
| |
Authors: | T. M. Carney R. M. Goldwyn |
| |
Affiliation: | (1) Department of Mechanical Engineering, Rice University, Houston, Texas;(2) NASA-Electronic Research Center, Cambridge, Massachusetts;(3) Present address: Rice University, Houston, Texas;(4) IBM-Watson Research Center, Yorktown Heights, New York |
| |
Abstract: | Numerical experiments are performed to measure the relative accuracy and computational efficiency of various estimators for the parameters and state of a linear dynamic system with forcing using a finite sequence of measurements containing noise. This nonlinear estimation problem is treated by estimators based on least-squares and maximumlikelihood criteria and, for the linearized problem, two mechanizations of the Kalman-Bucy estimator are applied. A digital computer simulation of an example problem is performed, and a Monte Carlo technique is used to generate statistics of the errors in the estimates empirically. This process is carried out for a range ofa priori error statistics.This work was supported by the National Science Foundation, Systems Grant No. GU-1153. |
| |
Keywords: | |
本文献已被 SpringerLink 等数据库收录! |
|