Last issue
2022 (vol. 32) - Number 1

J. Cvejn:

The magnitude optimum design of the PI controller for plants with complex roots and dead time

S.F. Al-Azzawi, M.A. Hayali:

Coexisting of self-excited and hidden attractors in a new 4D hyperchaotic Sprott-S system with a single equilibrium point

M.A. Hammami, N. El Houda Rettab, F. Delmotte:

On the state estimation for nonlinear continuous-time fuzzy systems

M. Ilyas, M.A. Khan, A. Khan, Wei Xie, Y. Khan:

Observer design estimating the propofol concentration in PKPD model with feedback control of anesthesia administration

L. Moysis, M. Tripathi, M. Marwan:

Adaptive observer design for systems with incremental quadratic constraints and nonlinear outputs – application to chaos synchronization

S. Vaidyanathan, K. Benkouider, A. Sambas:

A new multistable jerk chaotic system, its bifurcation analysis, backstepping control-based synchronization design and circuit simulation

T.T. Tuan, H. Zabiri, M.I.A. Mutalib, Dai-Viet N. Vo:

Disturbance-Kalman state for linear offset free MPC

Yuan Xu, Jun Wang:

A novel multiple attribute decision-making method based on Schweizer-Sklar t-norm and t-conorm with q-rung dual hesitant fuzzy information

T. Kaczorek:

Observers of fractional linear continuous-time systems

ACS Abstract:

2008 (Volume 18)
Number 3
1. Compensation of the scan-period irregularities in LQG control systems
2. Remarks about DC motor control
3. Concepts of learning in assembler encoding

Regulation of absorption packed column of CO2 using discrete fuzzy input-output linearization

5. Accelerator's supervisory control system based on CANbus
6. Stable output feedback model predictive control design: LMI approach

Compensation of the scan-period irregularities in LQG control systemsDownload full PDF article
Jan Cvejn
(University of Pardubice, Faculty of Electrical Engineering and Informatics, Czech Republic)

Computer-based control applications, especially if they run under general-purpose opera-ting systems, often exhibit variance of the scan period of processing inputs and outputs. Although this phenomenon is usually neglected when discrete control algorithms are used, it can cause worse performance of the control loop in comparison to theoretical case. In this paper we describe a modified discrete LQG control algorithm that takes disturbances of the scan period into account and partially compensates their effect. This modification concerns both the state estimation and generating the control output. We also show that such a controller can be implemented even on relatively simple hardware platforms if the system dynamics is time-invariant.

keywords: LQG controller, stochastic control, hybrid systems


Remarks about DC motor controlDownload full PDF article
Jerzy Baranowski, Marek Dlugosz, Wojciech Mitkowski
(Faculty of Electrical Engineering, AGH University of Science and Technology, Krakow, Poland)

This paper was motivated by known results in the electric motor control where classical control algorithms were verified in the experimental framework. We want to show, that the control theory offers more sophisticated control methods that await their practical implementation. Presented algorithms for both continuous and discrete control and state estimation give many possibilities of improvement in the field of electric drive control. Our results are illustrated with simulations, but with available technology (such as MATLAB/Simulink/RTWT) can be easily verified in practice.

keywords: DC motor control, linear-quadratic control, minimum-energy control, dead-beat control, dead-beat observer, discrete LQ control, nolinear observer, observer optimization, nonlinear dynamical feedback


Concepts of learning in assembler encodingDownload full PDF article
Tomasz Praczyk
(Naval University, Gdynia, Poland)

Assembler Encoding (AE) represents Artificial Neural Network (ANN) in the form of a simple program called Assembler Encoding Program (AEP). The task of AEP is to create the so-called Network Definition Matrix (NDM) maintaining the whole information necessary to construct ANN. To generate AEPs and in consequence ANNs genetic algorithms are used. Using evolution is one of the methods to create optimal ANNs. Another method is learning. During learning parameters of ANN, e.g. weights of interneuron connections, adjust to the task performed by ANN. Usually, combining both methods accelerates generating optimal ANNs. The paper addresses the problem of simultaneous use of the evolution and learning in AE.

keywords: evolutionary neural networks, reinforcement learning


Regulation of absorption packed column of CO2 using discrete fuzzy input-output linearization

Download full PDF article
R. Illoul, S. Bezzaoucha
(cole Polytechnique Nationale d'Alger, 10 Avenue Hacen Badi, El-Harrach, Algiers, Algeria)
A. Selatnia
(Laboratory of Process Control, Department of Chemical Engineering, Ecole Polytechnique Nationale d'Alger, Algiers, Algeria)

This work deals with modeling and regulation of the absorption packed column of CO2 from a gas mixture using an aqueous solution of monoethanolamine (MEA). We first present a dynamic mathematical model estimating the CO2 and MEA concentrations at column outlets; we compare the model results with the experimental data in the steady state case to validate the model. We first use the classical PI regulation technique to delete step input disturbances, then we present  the discrete fuzzy  input- output linearization theory that allows system inversion   and linearization for a large class of non linear systems and show that pole placement law guarantees error convergence to a reference centered ball. Finally we apply the regulator based on fuzzy linearization to the controlled system and we got in most cases better results than PI regulation.

keywords: CO2 absorption packed column, MEA, modeling, fuzzy control, PI regulation, input-output linearization


Accelerator's supervisory control system based on CANbusDownload full PDF article
Mirosław Dach
(PSI - Paul Scherrer Institut, Villigen, Switzerland)
Jan Werewka
(AGH University of Science and Technology, Department of Automatics, Computer Science Laboratory, Kraków, Poland)

This paper presents a structural approach to the Supervisory Control and Diagnostic System (SCDS) for highly distributed control systems. The Supervisory Control Systems are a class of control systems superintending the subordinate ones. The proposed methodology led to the construction of the low-cost SCDS system for the SLS (Swiss Light Source) accelerator project. In order to minimize the length of the accelerator's downtimes, it is essential to use the SCDS, which makes it possible to diagnose and react quickly to any failure of the accelerator's control system. During the realization of the Supervisory Control and Diagnostic System, it was considered to use CANbus with CANopen protocol and a PC running Linux with Real Time Application Interface linking the CANbus with Ethernet. To obtain a coherent concept of infrastructure, joining the constituent Supervisory Control System elements, the usefulness of object methodology for software creating was taken into account. The designed and constructed SCDS was subjected to a series of tests on its applicability to the planned tasks' execution. There were performed response time measurements for particular parts of the system which proved predefined assumptions. Thus, the accepted solutions were verified in practice.

keywords: RT systems, distributed systems, field bus, time analysis, CAN, accelerator


Stable output feedback model predictive control design: LMI approachDownload full PDF article
Vojtech Vesely
(URPI, Faculty of Electrical Engineering and IT, Slovak University of Technology, Bratislava, Slovak Republic)
Ruth Bars
(Budapest University of Technology and Economics, Department of Automation and Applied Information, Budapest, MTA-BME Control Research Group, Hungary)

The paper addresses two design problems dealing with a quadratic stable output/state feedback model  predictive control for linear systems without constraints. For the first approach the model predictive control is designed for a N2 state ahead prediction using  the Lyapunov function approach with cost function. In the second  approach,  the one step ahead prediction control is designed using classical LQR state feedback controller approach. By Diophantine matrix equation, the classical state feedback is recalculated to output one step ahead model predictive control.  Three examples are given to demonstrate the effectiveness of proposed methods.

keywords: model predictive control, quadratic stability, Lyapunov function


<< Back