Mihail Voicu - Academia.edu (original) (raw)

Papers by Mihail Voicu

IFAC Proceedings Volumes, Nov 1, 1985

The state and the oontrol of every spatial manipulating system are subject of certain presoribed ... more The state and the oontrol of every spatial manipulating system are subject of certain presoribed constraints. To satisfy these constraints the system must possess certain structural properties. The aim of this paper is to characterize such properties in connection with the state and control constraints, expressed in the form of time-dependent curved boxes in state space and in control space respectively. by using flow invariance methods. The obtained result may be used both for a more subtle evaluation of the dynamical evolution and for the design of the manipulation systems or of the control strategies of the robot systems.

IFAC Proceedings Volumes, Jun 1, 1994

Fuzzy and non-parametric theoretic-decision methods of pattern recognition are used for the model... more Fuzzy and non-parametric theoretic-decision methods of pattern recognition are used for the model based fault detection via parameter estimation. The present approach uses the concepts of fuzzy weighted distance and of total least-squares, on one hand, and the outputs of an optimized procedure of fuzzy clustering. on the other, to construct and to improve the performances of a linear trainable classifier of the least-squares minimum distance type. Simulation results regarding the diagnosis of a d.c. motor-pump-pipe process studied by means of numerical simulation are included.

IFAC Proceedings Volumes, Jul 1, 1984

Journal of Engineering Sciences and Innovation

Perron-Frobenius theory, well-known for non-negative matrices, is applied to the study of a categ... more Perron-Frobenius theory, well-known for non-negative matrices, is applied to the study of a category of economic systems described by Leontief-type models. For the considered models, the paper presents a qualitative point of view, oriented towards the solution existence / non-existence, as well as a quantitative point of view, oriented towards the solution construction. The main instruments used by our study are the structure of communication classes and the eigenvalue-eigenvector structure (also called matrix eigenstructure). A numerical case study is used for the illustration of theoretical concepts and results.

2022 26th International Conference on System Theory, Control and Computing (ICSTCC), Oct 19, 2022

2020 24th International Conference on System Theory, Control and Computing (ICSTCC), 2020

The current research extends the issue of componentwise stabilization of interval systems (ISs) -... more The current research extends the issue of componentwise stabilization of interval systems (ISs) - previously addressed by us, from a symmetrical to the broader, non-symmetrical framework. The considered issue means the individual monitoring of each closed-loop state variable by keeping it inside of a "box" that decreases exponentially. This approach relies on necessary and sufficient conditions for the flow-invariance of contractive rectangular sets with exponential decay. In the first part of the paper, two classes of feedback matrices ensuring componentwise stabilization are characterized as families of matrices solving some nonlinear algebraic inequalities; the larger matrix class corresponds to the general situation of the non-symmetrical invariant sets, which includes the particular case of symmetrical sets. The second part reformulates the results of the first one as linear programming problems, which guarantee the numerical tractability. Such a problem may be • feasible, when its solution provides a feedback matrix that ensures the componentwise stabilization, • unfeasible, when the interval system is not componentwise stabilizable. The third part studies the connections between the componentwise stability of the closed-loop IS and the standard stability properties, such as stability margin, existence of Lyapunov functions and Lyapunov inequalities. A numerical example illustrates the applicability of the technique proposed for IS componentwise stabilization.

2017 21st International Conference on System Theory, Control and Computing (ICSTCC), 2017

The paper provides new results for testing the flow-invariance of time-dependent sets with arbitr... more The paper provides new results for testing the flow-invariance of time-dependent sets with arbitrary shapes, with respect to trajectories of continuous-time nonlinear systems. The approach represents a generalization of a criterion used in the qualitative analysis of ordinary differential equations with solutions in closed sets with hyper-rectangular form. Given a nonlinear dynamical system for investigation, our generalization consists in the use of an auxiliary system with pseudo-linear form, whose trajectories play a comparison role for the trajectories of the investigated system. The various shapes of the invariant set candidates can be defined in terms of Hölder p-norms of some smooth time-dependent vector functions. The non-symmetrical and symmetrical invariant sets are addressed separately, since the construction of the auxiliary systems requires different properties for their trajectories. The developments we propose are able to include, as particular cases, several earlier studies on flow-invariance, which were separately reported in literature for different classes of dynamics.

2018 22nd International Conference on System Theory, Control and Computing (ICSTCC), 2018

The role played by column representatives is expanded in the study of the properties exhibited by... more The role played by column representatives is expanded in the study of the properties exhibited by the trajectories of arbitrary switching linear systems. Previous contributions on the employment of column representatives focused on positive dynamics and linear co positive Lyapunov functions associated with exponentially contractive positive sets that are invariant with respect to such dynamics. Our approach refers to arbitrary dynamics and invariant sets with general form for time-dependence. We address both discrete and continuous-time cases. Our key finding is that the existence of such invariant sets is fully characterized (if and only if) by the Schur (Hurwitz respectively) stability of the column representatives corresponding to a matrix set adequately built from the original system matrices. Our mathematical developments are illustrated by a numerical example. These developments incorporate the previous contributions mentioned above as particular cases.

2021 25th International Conference on System Theory, Control and Computing (ICSTCC), 2021

The paper constructs a mathematically rigorous support for flow-invariance analysis of discrete-t... more The paper constructs a mathematically rigorous support for flow-invariance analysis of discrete-time linear system dynamics, with respect to a large family of contractive sets, considered in the comprehensive form of proper C-sets, exponentially decreasing with a certain rate. The first part develops a general algebraic point of view for invariance characterization, based on an inequality built in terms of Minkowski functions (which incorporates the classical Stein-Lyapunov inequality, as a particular case). The second part shows that the inequality with a general form can be addressed by set-embedding techniques that offer scenarios with a better numerical tractability – stated as optimization problems. The role of these techniques is illustrated by the applications presented in the final part; the applications refer to two important classes of contractive sets (which were separately reported in literature, without any connection between the embedding principles), namely, polyhedral sets, and symmetrical sets with arbitrary shapes.

2019 23rd International Conference on System Theory, Control and Computing (ICSTCC), 2019

The work is placed within the area of qualitative analysis of arbitrary switching linear systems,... more The work is placed within the area of qualitative analysis of arbitrary switching linear systems, focusing on the existence of quadratic Lyapunov functions for discrete- or continuous-time dynamics. Our investigations express a series of connotations on the utilization of a criterion (amply cited in the control engineering literature) which characterizes the mode-dependent quadratic Lyapunov functions, in the discrete-time case. These connotations are formulated as two distinct theorems, for discrete-time and continuous-time systems, respectively; both theorems provide algebraic conditions that ensure the existence of time-independent quadratic Lyapunov functions, together with the associated sets invariant relative to system trajectories. In the discrete-time case, our algebraic condition includes (as a particular form) the algebraic condition used by the existing criterion mentioned above. Our theorems are accompanied by remarks that create a deeper insight into the dichotomy time-independence vs. time-dependence for quadratic Lyapunov functions. A numerical example addressed by the help of the SeDuMi solver with the YALMIP interface illustrates the key theoretical points discussed for continuous-time arbitrary switching systems.

IFAC Proceedings Volumes, 1998

lliMPUS forms part of the overall program PHARE initiated by the European Union for the economic ... more lliMPUS forms part of the overall program PHARE initiated by the European Union for the economic and social restructuring of the countries of Central and Eastern Europe. It has brought a historic contribution to University development in these countries, which represents a large-scale phenomenon with a considerable impact on the whole system of institutions and individuals engaged in Higher Education activity. The lliMPUS projects including Romanian partners support the national strategy promoted by the Ministry of Education for restructuring of the Higher Education in the spirit of internationalization and full academic recognition. The contribution to University development in the Central and Eastern European countries highlight the importance of pre-accession strategies to the European Union, implemented by lliMPUS-PHARE, as well as the future potential offered by SOCRAlliS-ERASMUS and LEONARDO DA VINCI programs.

IFAC Proceedings Volumes, 2001

This contribution presents an architecture for production planning and control, which combines an... more This contribution presents an architecture for production planning and control, which combines an artificial intelligence part, implemented as a multi-expert system with a simulation module. In this way the adaptability and flexibility requested by the present manufacturing companies can be achieved. The part which makes inferences, in a symbolic approach, is responsible for obtaining a schedule and control procedure; for evaluation and capacity planning assessment this knowledge based control system makes use of the simulation module. For this onc, it is easy to handle the temporal and quantitative aspects involved in flexible manufacturing planning.

at - Automatisierungstechnik, 1974

In der Arbeit wird eine graphisch-analytisch anwendbare Methode aufgezeigt, die es ermöglicht, Au... more In der Arbeit wird eine graphisch-analytisch anwendbare Methode aufgezeigt, die es ermöglicht, Aussagen über das Stabilitätsverhalten von Grenzschwingungen zu machen. Die auftretenden Grenzschwingungen selbst werden nach der Methode der harmonischen Balance [1] bestimmt

2015 19th International Conference on System Theory, Control and Computing (ICSTCC), 2015

Our paper proposes a framework that allows the construction of several types of diagonal Lyapunov... more Our paper proposes a framework that allows the construction of several types of diagonal Lyapunov functions for discrete- and continuous-time interval systems. The Lyapunov function candidates are defined as weighted vector norms, where the weighting matrices are diagonal and positive definite. For the existence of such Lyapunov functions, we provide necessary and sufficient conditions in the case of 1 and ℝ norms, as well as two sufficient conditions in the case of 2-norm. The results offer construction procedures focusing on diagonal Lyapunov functions with the fastest decreasing rate, fact that also ensures the identification of the fastest contractive invariant sets shaped in accordance with the aforementioned norms. The exposition covers the discrete- and continuous-time dynamics, in parallel, aiming to emphasize both similarities and differences between the two types of results. The applicability of the theoretical developments is illustrated by a numerical example.

Control Engineering Solutions: a practical approach

The control of thermal processes has been given much attention over recent years and the use of d... more The control of thermal processes has been given much attention over recent years and the use of digital controllers has been suggested for such real-life applications. These controllers often consist of discrete time versions of classical PID controllers or sometimes modern predictive and adaptive controllers. In this chapter a comparison is made between predictive and PID control, exemplified on a practical control problem.Temperature control of a thermal treatment furnace with PID and predictive algorithms, respectively, has been presented. Experimental results show a better behaviour from predictive control because of its ability to deal with time delay processes. It is worth noting that predictive control is strongly sensitive to the value of the weighting factor λ.

For a class of uncertain nonlinear systems (UNSs), the flow-invariance of a timedependent rectang... more For a class of uncertain nonlinear systems (UNSs), the flow-invariance of a timedependent rectangular set (TDRS) defines individual constraints for each component of the state-space trajectories. It is shown that the existence of the flowinvariance property is equivalent to the existence of positive solutions for some differential inequalities with constant coefficients (derived from the state-space equation of the UNS). Flow-invariance also provides basic tools for dealing with the componentwise asymptotic stability as a special type of asymptotic stability, where the evolution of the state variables approaching the equilibrium point (EP){0} is separately monitored (unlike the standard asymptotic stability, which relies on global information about the state variables, formulated in terms of norms). The EP{0} of a given UNS is proved to be componentwise asymptotically stable if and only if the EP{0} of a differential equation with constant coefficients is asymptotically stable in th...