Conference Program

Program of section Theory

Lectures

 

M. Stockmann, T. Friebel, R. Haber:
ROBUST AND FAST ESTIMATION OF COVARIANCE MATRIX
Presenter(s): Stockmann, Markus

Abstract

Nearly all multivariate methods (PCA, factor analysis, control chart, ...) for data analysis (outlier detection, fault detection, ...) require the estimation of covariance matrix. It will be shown, that based on the well-known parameters median and MAD (Median Absolute Deviation) a very fast and robust estimator of the covariance matrix can be achieved, which decreases significantly the sensitivity to outliers. This will be illustrated with calculations for principal component analysis and control charts. State of the art methods have two decisive disadvantages. First disadvantage is computational time. Second and more important disadvantage is that classical methods assume Gaussian distribution and robust estimation of covariance matrix for multivariate non-Gaussian distributed data fails. Via the new recommended method, robust estimation is also possible for non-Gaussian data. keywords: Covariance matrix, Robust estimation, Multivariate data analysis, PCA, Control charts

 

Aleksandra Nowacka-Leverton, Andrzej Bartoszewicz:
SWITCHING LINE DESIGN FOR VSC OF CABLE SUSPENDED LOAD WITH STATE CONSTRAINTS
Presenter(s): Not defined.

Abstract

The paper focuses on a new method of sliding mode control for cable suspended loads with acceleration and velocity constraints. The presented method employs a time-varying switching line which moves with a decreasing velocity and a constant angle of inclination to the origin of the error state space. Then, it stops moving and it remains fixed. Since at the initial time the line consists of these points in the error state space which represent initial conditions of the system, the reaching phase is eliminated. As a consequence the considered system is insensitive with respect to external disturbance and model uncertainty from the very beginning of the control action. In this paper, we consider a rope suspended mass m which is originally at rest at altitude x0 and is supposed to change its position to xd. Clearly, the mass cannot be required to move downwards with any acceleration exceeding gravity constant g and therefore in order to maintain some safety margin we assume that the mass acceleration cannot exceed amax, where amax is a constant smaller than g. Furthermore, the mass velocity is not allowed to be greater than vmax. In this paper we present the algorithm for selection of switching line parameters. In order to obtain fast system response, we design the sliding mode controller for the considered mass which ensures the minimum value of the introduced control quality criterion, i.e. the integral of the time multiplied by the absolute error (ITAE) and which guarantees that the specified constraints (the system acceleration and the system velocity) are satisfied during the whole control process. Moreover, the proposed method ensures that the tracking error converges to zero monotonically.

 

K.Miatliuk, F.Siemieniako:
DYNAMICS OF COORDINATION STRATEGIES OF HIERARCHICAL SYSTEMS
Presenter(s): Miatliuk, Kanstantsin

Abstract

The aim of the work is to present the coordination technology realized by coordinator of hierarchical systems in design and control tasks. the construction of hierarchical system by M.Mesarovich and Y. Takachara with its standard block (two-level system with dynamic objects – aed) has been chosen in the work as the theoretical basis. Aed (proposed by S.Novikava and K.Miatliuk) is a standard element of design systems, which realizes the general laws of systems organization on each level and the inter-level connections. The design and control tasks described above are performed in aed by coordinator on its three layers: choice, learning and self-organisation, by one of coordination strategy which corresponds to present state of the system organisation and coordinator information uncertainty. Coordination strategies are realized on choice layer. The change of coordinator states is executed on the learning and selforganization layers which are described by functions of states transition. As the result, the formal model of coordinator is given in the paper, coordination strategies and their dynamic in process of systems design and control are described, the exemplary coordination tasks are presented for the cases of the design of technological operations and biomechanical objects motion design and control.

 

Laviniu Bejenaru:
STUDY IN SIMULATED SPATIAL PROBLEM USING OPTIMAL CONTROL AND HYBRID SYSTEMS
Presenter(s): Bejenaru, Laviniu

Abstract

The study of practical problems involving hybrid systems is broadly used by control field. This paper presents an approach based on these types of systems and optimal control for solving a simulated situation rising from space technology field. This paper simulates a situation we are interested in of taking a known ground samples from an extraterrestrial body as Mars, dwarf planets etc. Our simulation stipulates that there exists a complete controlled spacecraft and his jobs is composed by three actions: to land in a safe mode, to take the ground samples in a number of predefined positions and, finally, to take-off safely and return toward Terra. Because the space missions are very complex, we present only a framework for this problem, without any attitude for an exhaustive solution. The problem is modeled by the hybrid automaton, involving the existence of possible hybrid states and the subsystems there are identified. The Pontryagin’s principle, in company with proper algorithms for the specific situation of some subsystems, is applied. The paper presents the solution and an example for the problem.

 

Víteček, A. - Tůma, J. - Vítečková, M.:
Active Control of Rotor Stability
Presenter(s): Not defined.

Abstract

The paper is devoted to active increasing of the rigid rotor stability in the journal bearing. The linear and nonlinear mathematical models of the rotor dynamics are considered. The results of the digital simulations and of the verification on the real model are presented.

 

Vítečková, M. - Víteček, A.:
Control of FOPTD Plants
Presenter(s): Not defined.

Abstract

The paper is devoted to the standard and 2DOF PI controller tuning for the FOPTD plants. Two tuning methods are derived in detail. Both methods are presented and compared. The use is shown on the examples.

 

Áron Ballagi and László T. Kóczy:
Loose cooperation of mobile robots
Presenter(s): Ballagi, Áron

Abstract

In control of autonomous mobile robots are many complex, well structured problems, where a hierarchical structure within the data is present. This means, that one or several components of the structure are determined at a higher level by a sub-tree of other components. The concept of fuzzy signatures was introduced to help model these kinds of problems. Fuzzy signatures which structure data into vectors of fuzzy values, each of which can be a further vector, handle complex structured data. This will widen the application of fuzzy theory to many areas where objects are complex and sometimes interdependent features are to be classified and similarities / dissimilarities evaluated. Often, human experts can and must make decisions based on comparisons of cases with different numbers of data components, with even some components missing. Fuzzy signature is created with this objective in mind. This tree structure is a generalization of fuzzy sets and vector valued fuzzy sets in a way modeling the human approach to complex problems. By examining the problem, an arbitrary structure belonging to the data set can be determined. Due to the fact that some components might be missing from a data element, the actual tree structures of the data may slightly differ. So that these data can be evaluated and compared, aggregation operators are given for each node in the arbitrary structure for the purpose of modifying the structure. To model problems with this type of dataset, fuzzy signature based rules can be constructed. Inferring a conclusion from such a model is a key issue. In this paper the formerly introduced idea of Mamdani-type inference in fuzzy signature based rule bases will be used for selecting records from an available data base which maximally match the requirements specified in a pattern. Finally a possible cooperative robot application on a realistic example with missing data components will be shown. The base idea of this example has come from the partly unpublished research projects at LIFE. The paper presents a loose cooperation system where a group of autonomous intelligent mobile robots is supposed to solve transportation problems according to the exact instruction given to the Robot Foreman (RF). The other robots have no direct communication links with RF and all others, but can solve the task by intention guessing from the actual movements and positions of other robots, even though they might not be unambiguous.

 

Dusan Krokavec, Anna Filasova:
State constrained control design
Presenter(s): Not defined.

Abstract

The paper presents a constructive method based on a classical memory-less feedback control for stabilization of discrete-time systems with equality constraints given on the state variables. The method ensures that the closed-loop system control law gain matrix is optimized while the matrix pseudoinverse initialization is used. The validity of the proposed method is demonstrated by a numerical example with an equality constraint tying together some of the state variables.

 

Anna Filasova, Dusan Krokavec:
LMI based design of virtual sensors
Presenter(s): Not defined.

Abstract

The main contribution of the paper is to present an LMI based design method for the virtual sensor design in continuous-time linear MIMO systems. Presented version is derived in the terms of optimization over LMI constraints where the standard LMI numerical optimization procedures are used to manipulate the estimator state-space description matrix parameters giving the stable estimator dynamics. The obtained formulation is a convex LMI problem for the full order estimators and handles desired optimal structure in continuous-time linear MIMO systems.

Posters