Algorithms for sustainable adaptive evaluation of the state of the stochastic control objects Текст научной статьи по специальности «Компьютерные и информационные науки»

3. ПоселовД.А. Информатика. Энциклопедический словарь. М.: Педагогика-Пресс, 1994. 648 с.

4. Официальный сайт издательства «Открытые Системы». Интернет университет информационных технологий. [Электронный ресурс]. Режим доступа: www.intuit.ru/ (дата обращения: 05.10.2010).

5. Официальный русскоязычный сайт разработчиков WikiMedia. Интернет-энциклопедия Wikipedia. [Электронный ресурс]. Режим доступа: ru.wikipedia.org/wiki/Численное_моделирование/ (дата обращения: 5.10.2010).

ALGORITHMS FOR SUSTAINABLE ADAPTIVE EVALUATION OF THE STATE OF THE STOCHASTIC CONTROL OBJECTS Kodirov D.T. (Republic of Uzbekistan) Email: [email protected]

Kodirov Dilmurod Tukhtasunovich — Applicant, DEPARTMENT INFORMATION PROCESSING SYSTEMS AND MANAGEMENT, TASHKENT STATE TECHNICAL UNIVERSITY, TASHKENT, REPUBLIC OF UZBEKISTAN

Abstract: the problems of constructing algorithms for sustainable adaptive estimation of the state of stochastic control objects are considered. To improve the quality of filtering, regular adaptive assessment methods are used. These algorithms allow synthesized algorithms for adaptive estimation of the state of controlled objects in the presence of uncertain parameters, which allow decomposing the estimation problem, increasing the computational stability of algorithms, and reducing the amount of computational costs. Keywords: state of a stochastic control object, adaptive estimation, stable algorithms.

АЛГОРИТМЫ УСТОЙЧИВОГО АДАПТИВНОГО ОЦЕНИВАНИЯ СОСТОЯНИЯ СТОХАСТИЧЕСКИХ ОБЪЕКТОВ УПРАВЛЕНИЯ Кодиров Д.Т. (Республика Узбекистан)

Кодиров Дилмурод Тухтасунович — соискатель, кафедра систем обработки информации и управления, Ташкентский государственный технический университет, г. Ташкент, Республика Узбекистан

Аннотация: рассматриваются вопросы построения алгоритмов устойчивого адаптивного оценивания состояния стохастических объектов управления. Для повышения качества фильтрации используются методы регулярного адаптивного оценивания. Приведенные алгоритмы позволяют синтезировть алгоритмы адаптивного оценивания состояния управляемых объектов при наличии неопределенных параметров, позволяющих декомпозировать задачу оценивания, повысить вычислительную устойчивость алгоритмов, и сократить объем вычислительных затрат. Ключевые слова: состояние стохастического объекта управления, адаптивное оценивание, устойчивые алгоритмы.

The theory of optimum filtration is more fully developed now for conditions when statistical characteristics of signals and hindrances at creation of the filter are known precisely. However in practice at the solution of specific objectives of automation and management of technological objects these conditions often are not satisfied as there is an aprioristic uncertainty of characteristics of messages and hindrances. It interferes with creation of optimum filters and achievement of extreme accuracy of filtration. In literature the set of approaches to creation of adaptive filters and various algorithms of adaptive filtration received on their basis are described [1, 2]. One of the most effective methods of synthesis of adaptive and suboptimal filters is based on estimation methods with use of Gaussian approximation of a posteriori density of probability of a required vector of a state. However at practical applications the quality of filtration brought by these algorithms not always is acceptable. It is caused by the fact that at a task of the model of an object of management used in a filtration algorithm inevitably there are mistakes. For overcoming action of the specified reasons there is a need of development of effective modifications of the recurrent of the filters Kalman type answering, in particular, to regularization methods incorrectly of objectives. In this regard development of algorithms of regular adaptive estimation of a condition of stochastic objects of management on the basis of methods of the theory of conditional and Gaussian filtration is represented relevant.

In the report the synthesis algorithms of the adaptive filter based on the identifications of unknown average values and dispersion of noise of model of dynamics allowing to increase efficiency of modeling and the solution of problems of forecasting on the basis of dynamic of the filters kalman type are offered. The regular algorithms of adaptation of covariation matrixes of noise of an object and measurements by the principle of a maximum of credibility allowing to carry out effective estimation of a condition of a linear dynamic system in the conditions of the changing characteristics of noise of process and measurements are developed. The regular algorithms of steady adaptive estimation of a condition of the operated objects in the conditions of aprioristic uncertainty which are not demanding full a priori knowledge of an object and conditions of its functioning are offered. The algorithms of regular estimation of an initial condition of objects of management in the conditions of the limited volume of initial information allowing to estimate more more precisely admissible limits of inaccuracy of a task of a calculated value of a covariation matrix of an error of assessment of an initial state are developed. The algorithms of adaptive estimation in the presence of uncertain parameters allowing to dekompozirovat a problem of estimation, to increase computing stability of algorithms, and to reduce the volume of computing expenses are synthesized.

References / Список литературы

1. Sinitsin I.N. Filtri Kalmana i Pugacheva. Izd-vo: Logos, 2006. 640 s.

2. Igamberdiyev X.Z., Yusupbekov A.N., Zaripov O.O. Regulyarniye metodi otsenivaniya i upravleniya dinamicheskimi obyektami v usloviyax neopredelennosti. T.: TashGTU, 2012. 320 s.

SYNTHESIS ALGORITHMS OF OBSERVERS IN CONTROL SYSTEMS

OF DYNAMIC OBJECTS Rasulev A.Kh. (Republic of Uzbekistan) Email: [email protected]

Rasulev Aliakbar Khamidullayevich — Applicant, DEPARTMENT INFORMATION PROCESSING SYSTEMS AND MANAGEMENT, TASHKENT STATE TECHNICAL UNIVERSITY, TASHKENT, REPUBLIC OF UZBEKISTAN

Abstract: in the report questions of synthesis of observers in control systems of dynamic objects are considered. Synthesis algorithms of observers of indignations in control systems of dynamic objects of the full and lowered orders are offered. The given algorithms allow to solve effectively a problem of synthesis of invariant control systems of dynamic objects in case of external influences inaccessible to direct measurement. The offered algorithms are used at the solution of a problem of automation ofproduction control of adsorptive division of gas mix.

Keywords: dynamic objects, control systems, synthesis algorithms of observers.

АЛГОРИТМЫ СИНТЕЗА НАБЛЮДАТЕЛЕЙ В СИСТЕМАХ УПРАВЛЕНИЯ ДИНАМИЧЕСКИМИ ОБЪЕКТАМИ Расулев А.Х. (Республика Узбекистан)

Расулев Алиакбар Хамидуллаевич — соискатель, кафедра систем обработки информации и управления, Ташкентский государственный технический университет, г. Ташкент, Республика Узбекистан

Аннотация: в докладе рассматриваются вопросы синтеза наблюдателей в системах управления динамическими объектами. Предложены алгоритмы синтеза наблюдателей возмущений в системах управления динамическими объектами полного и пониженного порядков. Приведенные алгоритмы позволяют эффективно решать задачу синтеза инвариантных систем управления динамическими объектами в случае недоступных прямому измерению внешних воздействий. Предложенные алгоритмы использованы при решении задачи автоматизации управления технологическим процессом адсорбционного разделения газовой смеси.

Ключевые слова: динамические объекты, системы управления, алгоритмы синтеза наблюдателей.

Observing devices of external influences are used in control devices of combined and invariant automatic systems. This allows us to significantly improve the properties of the latter, both in transitional and steady-state modes [1, 2]. When choosing the gain of the observer, you should strive to ensure that the dynamic

26