Differential neural networks observer for second order systems with sampled and quantized output

Angel Avelar, Ivan Salgado, Hafiz Ahmed, Manuel Mera, Issac Chairez

Research output: Contribution to journalArticle

Abstract

Real instrumentation of control systems in digital devices introduces the necessity ofconsidering quantization and sampling information used in the control and estimator design. The aimof this study is designing a state estimator for uncertain second order nonlinear systems based onthe approximation enforced by differential neural networks (DNN) with quantized and time-varyingsampled output information. The effect of sampling output information is considered as a time-varyingdelay. The DNN estimates the set of non-linearities in the system structure with the applications of anadaptive approximation. A Lyapunov-Krasovskii functional served to justify the design of the law thatadjusted the weights of the DNN. The origin of the estimation error space is practically stable with theapproximation enforced by the DNN. Experimental results implement the DNN observer to reconstructthe states of the Van Der Pol Oscillator. The estimation attained with the proposed observer is comparedwith the results provided by classical linear observer. The evaluation of the least mean square errordemonstrates the superior performance of the solution suggested in this study. The Lyapunov-Krasovskiimethodology estimates the region of convergence depending on the sampled period and the level ofquantization.
LanguageEnglish
Pages490-495
Number of pages6
JournalIFAC-PapersOnLine
Volume51
Issue number3
DOIs
Publication statusPublished - 31 Aug 2018

Fingerprint

Neural networks
Sampling
Digital devices
Error analysis
Nonlinear systems
Control systems

Cite this

Differential neural networks observer for second order systems with sampled and quantized output. / Avelar, Angel ; Salgado, Ivan ; Ahmed, Hafiz; Mera, Manuel; Chairez, Issac .

In: IFAC-PapersOnLine, Vol. 51, No. 3, 31.08.2018, p. 490-495.

Research output: Contribution to journalArticle

Avelar, Angel ; Salgado, Ivan ; Ahmed, Hafiz ; Mera, Manuel ; Chairez, Issac . / Differential neural networks observer for second order systems with sampled and quantized output. In: IFAC-PapersOnLine. 2018 ; Vol. 51, No. 3. pp. 490-495.
@article{a15d6df7c9084a9c9c70b107875a6721,
title = "Differential neural networks observer for second order systems with sampled and quantized output",
abstract = "Real instrumentation of control systems in digital devices introduces the necessity ofconsidering quantization and sampling information used in the control and estimator design. The aimof this study is designing a state estimator for uncertain second order nonlinear systems based onthe approximation enforced by differential neural networks (DNN) with quantized and time-varyingsampled output information. The effect of sampling output information is considered as a time-varyingdelay. The DNN estimates the set of non-linearities in the system structure with the applications of anadaptive approximation. A Lyapunov-Krasovskii functional served to justify the design of the law thatadjusted the weights of the DNN. The origin of the estimation error space is practically stable with theapproximation enforced by the DNN. Experimental results implement the DNN observer to reconstructthe states of the Van Der Pol Oscillator. The estimation attained with the proposed observer is comparedwith the results provided by classical linear observer. The evaluation of the least mean square errordemonstrates the superior performance of the solution suggested in this study. The Lyapunov-Krasovskiimethodology estimates the region of convergence depending on the sampled period and the level ofquantization.",
author = "Angel Avelar and Ivan Salgado and Hafiz Ahmed and Manuel Mera and Issac Chairez",
year = "2018",
month = "8",
day = "31",
doi = "10.1016/j.ifacol.2018.07.327",
language = "English",
volume = "51",
pages = "490--495",
journal = "IFAC-PapersOnLine",
issn = "2405-8963",
publisher = "IFAC Secretariat",
number = "3",

}

TY - JOUR

T1 - Differential neural networks observer for second order systems with sampled and quantized output

AU - Avelar, Angel

AU - Salgado, Ivan

AU - Ahmed, Hafiz

AU - Mera, Manuel

AU - Chairez, Issac

PY - 2018/8/31

Y1 - 2018/8/31

N2 - Real instrumentation of control systems in digital devices introduces the necessity ofconsidering quantization and sampling information used in the control and estimator design. The aimof this study is designing a state estimator for uncertain second order nonlinear systems based onthe approximation enforced by differential neural networks (DNN) with quantized and time-varyingsampled output information. The effect of sampling output information is considered as a time-varyingdelay. The DNN estimates the set of non-linearities in the system structure with the applications of anadaptive approximation. A Lyapunov-Krasovskii functional served to justify the design of the law thatadjusted the weights of the DNN. The origin of the estimation error space is practically stable with theapproximation enforced by the DNN. Experimental results implement the DNN observer to reconstructthe states of the Van Der Pol Oscillator. The estimation attained with the proposed observer is comparedwith the results provided by classical linear observer. The evaluation of the least mean square errordemonstrates the superior performance of the solution suggested in this study. The Lyapunov-Krasovskiimethodology estimates the region of convergence depending on the sampled period and the level ofquantization.

AB - Real instrumentation of control systems in digital devices introduces the necessity ofconsidering quantization and sampling information used in the control and estimator design. The aimof this study is designing a state estimator for uncertain second order nonlinear systems based onthe approximation enforced by differential neural networks (DNN) with quantized and time-varyingsampled output information. The effect of sampling output information is considered as a time-varyingdelay. The DNN estimates the set of non-linearities in the system structure with the applications of anadaptive approximation. A Lyapunov-Krasovskii functional served to justify the design of the law thatadjusted the weights of the DNN. The origin of the estimation error space is practically stable with theapproximation enforced by the DNN. Experimental results implement the DNN observer to reconstructthe states of the Van Der Pol Oscillator. The estimation attained with the proposed observer is comparedwith the results provided by classical linear observer. The evaluation of the least mean square errordemonstrates the superior performance of the solution suggested in this study. The Lyapunov-Krasovskiimethodology estimates the region of convergence depending on the sampled period and the level ofquantization.

U2 - 10.1016/j.ifacol.2018.07.327

DO - 10.1016/j.ifacol.2018.07.327

M3 - Article

VL - 51

SP - 490

EP - 495

JO - IFAC-PapersOnLine

T2 - IFAC-PapersOnLine

JF - IFAC-PapersOnLine

SN - 2405-8963

IS - 3

ER -