New generalized least squares estimation algorithm for system identification

1 December 1973

conference paper
Published by Institute of Electrical and Electronics Engineers (IEEE)

Vol. 12, 755-759
https://doi.org/10.1109/cdc.1973.269112

Abstract

This paper presents modified formulations of the generalized least squares estimation algorithm for system parameter identification. Two sets of results are derived: First the existing algorithm is reformulated by eliminating the intermediate filtering procedures and introducing the system's input-output correlation matrices. Then it is shown that the new algorithm can be conveniently simplified to a form which does not require repeated matrix inversions.

Keywords

LEAST SQUARES APPROXIMATION
SYSTEM IDENTIFICATION
PARAMETER ESTIMATION
EQUATIONS
FILTERING ALGORITHMS
RANDOM SEQUENCES
TRANSFER FUNCTIONS
FILTERS
LEAST SQUARES METHODS

This publication has 0 references indexed in Scilit: