Publication result detail

Adaptive Controllers with Process Identification Based on Modern Control Methods

VAŇKOVÁ, M.

Original Title

Adaptive Controllers with Process Identification Based on Modern Control Methods

English Title

Adaptive Controllers with Process Identification Based on Modern Control Methods

Type

Paper in proceedings (conference paper)

Original Abstract

The paper deals with problems connected with process identification. The process model estimation is the main part of adaptive controller. Classical recursive identification algorithms (RLS, LDFIL) are discussed. The neural network based identification using backpropagation and Marquardt-Levenberg training algorithm is also introduced as a solution of defects of the classical identification algorithm. The identification algorithms are compared in practical example.

English abstract

The paper deals with problems connected with process identification. The process model estimation is the main part of adaptive controller. Classical recursive identification algorithms (RLS, LDFIL) are discussed. The neural network based identification using backpropagation and Marquardt-Levenberg training algorithm is also introduced as a solution of defects of the classical identification algorithm. The identification algorithms are compared in practical example.

Key words in English

adaptive control, recursive algorithm, neural network identification, quantization effect

Authors

VAŇKOVÁ, M.

Released

01.01.2006

Publisher

Univerzita Pardubice

Location

Pardubice

ISBN

80-7194-860-8

Book

Proceedings the 7th International Scientific-Technical Conference PROCESS CONTROL 2006

Pages from

54

Pages count

6

BibTex

@inproceedings{BUT19414,
  author="Markéta {Šulová}",
  title="Adaptive Controllers with Process Identification Based on Modern Control Methods",
  booktitle="Proceedings the 7th International Scientific-Technical Conference PROCESS CONTROL 2006",
  year="2006",
  pages="6",
  publisher="Univerzita Pardubice",
  address="Pardubice",
  isbn="80-7194-860-8"
}