A model-free design of reduced-order controllers and application to a DC servomotor

Sofiane Khadraoui, Hazem Nounou, Mohamed Nounou, Aniruddha Datta, Shankar P. Bhattacharyya

Research output: Contribution to journalArticle

12 Citations (Scopus)

Abstract

This paper presents a new model-free technique to design fixed-structure controllers for linear unknown systems. In the current control design approaches, measured data are used to first identify a model of the plant, then a controller is designed based on the identified model. Due to errors associated with the identification process, degradation in the controller performance is expected. Hence, we use the measured data to directly design the controller without the need for model identification. Our objective here is to design measurement-based controllers for stable and unstable systems, even when the closed-loop architecture is unknown. This proposed method can be very useful for many industrial applications. The proposed control methodology is a reference model design approach which aims at finding suitable parameter values of a fixed-order controller so that the closed-loop frequency response matches a desired frequency response. This reference model design problem is formulated as a nonlinear programming problem using the concept of bounded error, which can then be solved to find suitable values of the controller parameters. In addition to the well-known advantages of data-based control methods, the main features of our proposed approach are: (1) the error is guaranteed to be bounded, (2) it enables us to avoid issues related to the use of minimization methods, (3) it can be applied to stable and unstable plants and does not require any knowledge about the closed-loop architecture, and (4) the controller structure can be selected a priori, which means that low-order controllers can be designed. The proposed technique is experimentally validated through a real position control problem of a DC servomotor, where the results demonstrate the efficacy of the proposed method.

Original languageEnglish
Pages (from-to)2142-2149
Number of pages8
JournalAutomatica
Volume50
Issue number8
DOIs
Publication statusPublished - 2014

Fingerprint

Servomotors
Controllers
Frequency response
Identification (control systems)
Position control
Electric current control
Nonlinear programming
Industrial applications
Linear systems
Degradation

Keywords

  • Control of DC servomotor
  • Controller tuning
  • Frequency response
  • Low-order control design
  • Model-free control
  • Performance achievement
  • Reference model design
  • Unknown control architecture

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Electrical and Electronic Engineering

Cite this

A model-free design of reduced-order controllers and application to a DC servomotor. / Khadraoui, Sofiane; Nounou, Hazem; Nounou, Mohamed; Datta, Aniruddha; Bhattacharyya, Shankar P.

In: Automatica, Vol. 50, No. 8, 2014, p. 2142-2149.

Research output: Contribution to journalArticle

Khadraoui, Sofiane ; Nounou, Hazem ; Nounou, Mohamed ; Datta, Aniruddha ; Bhattacharyya, Shankar P. / A model-free design of reduced-order controllers and application to a DC servomotor. In: Automatica. 2014 ; Vol. 50, No. 8. pp. 2142-2149.
@article{69e5e3fb606e4b4ea7fd110b7cf83ab3,
title = "A model-free design of reduced-order controllers and application to a DC servomotor",
abstract = "This paper presents a new model-free technique to design fixed-structure controllers for linear unknown systems. In the current control design approaches, measured data are used to first identify a model of the plant, then a controller is designed based on the identified model. Due to errors associated with the identification process, degradation in the controller performance is expected. Hence, we use the measured data to directly design the controller without the need for model identification. Our objective here is to design measurement-based controllers for stable and unstable systems, even when the closed-loop architecture is unknown. This proposed method can be very useful for many industrial applications. The proposed control methodology is a reference model design approach which aims at finding suitable parameter values of a fixed-order controller so that the closed-loop frequency response matches a desired frequency response. This reference model design problem is formulated as a nonlinear programming problem using the concept of bounded error, which can then be solved to find suitable values of the controller parameters. In addition to the well-known advantages of data-based control methods, the main features of our proposed approach are: (1) the error is guaranteed to be bounded, (2) it enables us to avoid issues related to the use of minimization methods, (3) it can be applied to stable and unstable plants and does not require any knowledge about the closed-loop architecture, and (4) the controller structure can be selected a priori, which means that low-order controllers can be designed. The proposed technique is experimentally validated through a real position control problem of a DC servomotor, where the results demonstrate the efficacy of the proposed method.",
keywords = "Control of DC servomotor, Controller tuning, Frequency response, Low-order control design, Model-free control, Performance achievement, Reference model design, Unknown control architecture",
author = "Sofiane Khadraoui and Hazem Nounou and Mohamed Nounou and Aniruddha Datta and Bhattacharyya, {Shankar P.}",
year = "2014",
doi = "10.1016/j.automatica.2014.06.001",
language = "English",
volume = "50",
pages = "2142--2149",
journal = "Automatica",
issn = "0005-1098",
publisher = "Elsevier Limited",
number = "8",

}

TY - JOUR

T1 - A model-free design of reduced-order controllers and application to a DC servomotor

AU - Khadraoui, Sofiane

AU - Nounou, Hazem

AU - Nounou, Mohamed

AU - Datta, Aniruddha

AU - Bhattacharyya, Shankar P.

PY - 2014

Y1 - 2014

N2 - This paper presents a new model-free technique to design fixed-structure controllers for linear unknown systems. In the current control design approaches, measured data are used to first identify a model of the plant, then a controller is designed based on the identified model. Due to errors associated with the identification process, degradation in the controller performance is expected. Hence, we use the measured data to directly design the controller without the need for model identification. Our objective here is to design measurement-based controllers for stable and unstable systems, even when the closed-loop architecture is unknown. This proposed method can be very useful for many industrial applications. The proposed control methodology is a reference model design approach which aims at finding suitable parameter values of a fixed-order controller so that the closed-loop frequency response matches a desired frequency response. This reference model design problem is formulated as a nonlinear programming problem using the concept of bounded error, which can then be solved to find suitable values of the controller parameters. In addition to the well-known advantages of data-based control methods, the main features of our proposed approach are: (1) the error is guaranteed to be bounded, (2) it enables us to avoid issues related to the use of minimization methods, (3) it can be applied to stable and unstable plants and does not require any knowledge about the closed-loop architecture, and (4) the controller structure can be selected a priori, which means that low-order controllers can be designed. The proposed technique is experimentally validated through a real position control problem of a DC servomotor, where the results demonstrate the efficacy of the proposed method.

AB - This paper presents a new model-free technique to design fixed-structure controllers for linear unknown systems. In the current control design approaches, measured data are used to first identify a model of the plant, then a controller is designed based on the identified model. Due to errors associated with the identification process, degradation in the controller performance is expected. Hence, we use the measured data to directly design the controller without the need for model identification. Our objective here is to design measurement-based controllers for stable and unstable systems, even when the closed-loop architecture is unknown. This proposed method can be very useful for many industrial applications. The proposed control methodology is a reference model design approach which aims at finding suitable parameter values of a fixed-order controller so that the closed-loop frequency response matches a desired frequency response. This reference model design problem is formulated as a nonlinear programming problem using the concept of bounded error, which can then be solved to find suitable values of the controller parameters. In addition to the well-known advantages of data-based control methods, the main features of our proposed approach are: (1) the error is guaranteed to be bounded, (2) it enables us to avoid issues related to the use of minimization methods, (3) it can be applied to stable and unstable plants and does not require any knowledge about the closed-loop architecture, and (4) the controller structure can be selected a priori, which means that low-order controllers can be designed. The proposed technique is experimentally validated through a real position control problem of a DC servomotor, where the results demonstrate the efficacy of the proposed method.

KW - Control of DC servomotor

KW - Controller tuning

KW - Frequency response

KW - Low-order control design

KW - Model-free control

KW - Performance achievement

KW - Reference model design

KW - Unknown control architecture

UR - http://www.scopus.com/inward/record.url?scp=84905261545&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84905261545&partnerID=8YFLogxK

U2 - 10.1016/j.automatica.2014.06.001

DO - 10.1016/j.automatica.2014.06.001

M3 - Article

AN - SCOPUS:84905261545

VL - 50

SP - 2142

EP - 2149

JO - Automatica

JF - Automatica

SN - 0005-1098

IS - 8

ER -