Advanced

Robust On-Line Estimation

Gustafsson, Lars and Olsson, Mikael (1999) In MSc Theses
Department of Automatic Control
Abstract
In the presence if poor excitation and abrupt changes of a time-varying system the Recursive least square (RLS) with forgetting factor is not able to track the parameters in a suitable way. The purpose of this thesis is to investigate different kinds of proposals of algorithms to deal with the phenomenons arising, such as estimator windup or too slow convergence of the estimations. Attention is also paid to prior information of the bounds of the system and two variants of RLS are presented to prevent the estimation to exceed these bounds. All algorithms are introduced theoretically and their performances are verified via simulation studies. <br><br> Finally the application of the algorithms is illustrated. Incorporating four of our... (More)
In the presence if poor excitation and abrupt changes of a time-varying system the Recursive least square (RLS) with forgetting factor is not able to track the parameters in a suitable way. The purpose of this thesis is to investigate different kinds of proposals of algorithms to deal with the phenomenons arising, such as estimator windup or too slow convergence of the estimations. Attention is also paid to prior information of the bounds of the system and two variants of RLS are presented to prevent the estimation to exceed these bounds. All algorithms are introduced theoretically and their performances are verified via simulation studies. <br><br> Finally the application of the algorithms is illustrated. Incorporating four of our algorithms in the RLD we succeed in estimating the mass of a car and the slope of a road on-line by simulating a mathematical model. (Less)
Please use this url to cite or link to this publication:
author
Gustafsson, Lars and Olsson, Mikael
supervisor
organization
year
type
H3 - Professional qualifications (4 Years - )
subject
publication/series
MSc Theses
report number
TFRT-5633
ISSN
0280-5316
language
English
id
8848504
date added to LUP
2016-03-24 11:15:56
date last changed
2016-03-24 11:15:56
@misc{8848504,
  abstract     = {In the presence if poor excitation and abrupt changes of a time-varying system the Recursive least square (RLS) with forgetting factor is not able to track the parameters in a suitable way. The purpose of this thesis is to investigate different kinds of proposals of algorithms to deal with the phenomenons arising, such as estimator windup or too slow convergence of the estimations. Attention is also paid to prior information of the bounds of the system and two variants of RLS are presented to prevent the estimation to exceed these bounds. All algorithms are introduced theoretically and their performances are verified via simulation studies. <br><br> Finally the application of the algorithms is illustrated. Incorporating four of our algorithms in the RLD we succeed in estimating the mass of a car and the slope of a road on-line by simulating a mathematical model.},
  author       = {Gustafsson, Lars and Olsson, Mikael},
  issn         = {0280-5316},
  language     = {eng},
  note         = {Student Paper},
  series       = {MSc Theses},
  title        = {Robust On-Line Estimation},
  year         = {1999},
}