NONLINEAR REGRESSION MODEL PARAMETER ESTIMATION THROUGH RANDOM ORDERING OF MEASUREMENTS.

T. G. Robertazzi; S. C. Schwartz

NONLINEAR REGRESSION MODEL PARAMETER ESTIMATION THROUGH RANDOM ORDERING OF MEASUREMENTS.

T. G. Robertazzi
, S. C. Schwartz

Electrical Engineering

Research output: Contribution to conference › Paper › peer-review

1 Scopus citations

Abstract

An extended Kalman filter is prone to diverge if applied to a nonlinear regression model where minimizing the error over a local portion of the data does not ensure that it is minimized globally. Processing the measurements in a random order, rather than in the causal one in which they occur, can avoid this problem. This is illustrated through simulations involving a particular model.

Original language	English
Pages	946-954
Number of pages	9
State	Published - 1980
Event	Proc Annu Allerton Conf Commun Control Comput 18th - Monticello, IL, USA Duration: Oct 8 1980 → Oct 11 1980

Conference

Conference	Proc Annu Allerton Conf Commun Control Comput 18th
City	Monticello, IL, USA
Period	10/8/80 → 10/11/80

Cite this

@conference{48eb245dcb9c480b8727fc4adb6270b1,

title = "NONLINEAR REGRESSION MODEL PARAMETER ESTIMATION THROUGH RANDOM ORDERING OF MEASUREMENTS.",

abstract = "An extended Kalman filter is prone to diverge if applied to a nonlinear regression model where minimizing the error over a local portion of the data does not ensure that it is minimized globally. Processing the measurements in a random order, rather than in the causal one in which they occur, can avoid this problem. This is illustrated through simulations involving a particular model.",

author = "Robertazzi, \{T. G.\} and Schwartz, \{S. C.\}",

year = "1980",

language = "English",

pages = "946--954",

note = "Proc Annu Allerton Conf Commun Control Comput 18th ; Conference date: 08-10-1980 Through 11-10-1980",

}

TY - CONF

T1 - NONLINEAR REGRESSION MODEL PARAMETER ESTIMATION THROUGH RANDOM ORDERING OF MEASUREMENTS.

AU - Robertazzi, T. G.

AU - Schwartz, S. C.

PY - 1980

Y1 - 1980

N2 - An extended Kalman filter is prone to diverge if applied to a nonlinear regression model where minimizing the error over a local portion of the data does not ensure that it is minimized globally. Processing the measurements in a random order, rather than in the causal one in which they occur, can avoid this problem. This is illustrated through simulations involving a particular model.

AB - An extended Kalman filter is prone to diverge if applied to a nonlinear regression model where minimizing the error over a local portion of the data does not ensure that it is minimized globally. Processing the measurements in a random order, rather than in the causal one in which they occur, can avoid this problem. This is illustrated through simulations involving a particular model.

UR - https://www.scopus.com/pages/publications/84941604056

M3 - Paper

AN - SCOPUS:84941604056

SP - 946

EP - 954

T2 - Proc Annu Allerton Conf Commun Control Comput 18th

Y2 - 8 October 1980 through 11 October 1980

ER -

NONLINEAR REGRESSION MODEL PARAMETER ESTIMATION THROUGH RANDOM ORDERING OF MEASUREMENTS.

Abstract

Conference

Other files and links

Fingerprint

Cite this