Gateway for:

Member Countries

Self-Organization Processes in Ecological Applications

#B-477


Development of Self-Organization Methods in Complex Systems for Exploring Problems of Physical Structure Formation and Signal Recognition in Ecological Monitoring Tasks

Tech Area / Field

  • ENV-APC/Air Pollution and Control/Environment
  • ENV-MIN/Monitoring and Instrumentation/Environment
  • INF-SIG/Sensors and Signal Processing/Information and Communications
  • PHY-NGD/Fluid Mechanics and Gas Dynamics/Physics
  • PHY-OPL/Optics and Lasers/Physics

Status
3 Approved without Funding

Registration date
29.02.2000

Leading Institute
National Academy of Sciences of the Republic of Belarus / Institute of Informatics Problems, Belarus, Minsk

Supporting institutes

  • Military Academy of Belarus, Belarus, Minsk\nBelarussian State University, Belarus, Minsk\nB.I. Stepanov Institute of Physics, Belarus, Minsk

Collaborators

  • University of Florence / Department of Physics, Italy, Florence

Project summary

The project purpose. The state of the art and the impact of the project on the scientific and engineering progress. The purpose of the present project is the development of self-organization methods in complex systems for exploring problems of physical structures formation and signals recognition in ecological monitoring tasks.

The problems of environmental monitoring are that area, where the application of methods and theory of complex dynamic systems has the objective promises to give fast and significant result. The creation of global and regional monitoring systems of atmosphere, ocean, Earth surface has become the actual task, in the solution of which a scientific and technical potential of world community is integrated.

In the project the developed methods of the description of complex dynamic systems will be used for development of the new approaches to the solution of the main tasks in data processing of remote probing. One of the fields of application of the theory of complex dynamic systems to the tasks of monitoring of an environment will be retrieving of aerosol atmospheric component by results of laser sounding. The transformation of an aerosol atmospheric layer exerts direct influence on the formation of an Earth climate that is the important factor determining formation of clouds, dynamics of an ozone layer - the derivation of ecological dangerous situations. The lidars are the most effective equipment of remote probing of atmospheric components including aerosols. The aerosol concentration data and aerosol microstructure data are retrieved from the data of a multiwave laser probing at simultaneous registration of the polarization characteristics of location signal.

As concrete object of researches in the project two natural systems (media) are investigated: the stratosphere aerosol layer representing a global system, the changes in which happen on distances in hundreds and thousands kilometers during from several weeks till several years, and the aerosol clouds derivated as a result of technogenic emission in low layers of atmosphere.

Within the framework of the present project it is being planned to offer models of investigated objects with using complex systems equations, and also algorithms, developed in B-95 ISTC project, for digital processing of experimental ecological monitoring data on the basis of methods of complex systems analysis. Recently in development of the theory and methods of self-organization in complex systems some paradigms and conceptual models were laid down. As the Nobel prize winner I.Prigogine has noted, real or physical (with irreversible time) the description of complex systems and processes generated by them " is suit only for some types of dynamic systems ". In particular, on the one hand, it takes place for dynamic systems with deterministic chaos or chaotic dynamic systems (CDS), and on the other hand, for so-called large Poincare systems (LPS), i.e. for systems consisting of large number of permanently interacting particles. A reason of origin of time irreversibility in LPS is quite different than in case of CDS. In connection with LPS refer to nonintegrable by Poincare class of systems (i.e. systems, in which interaction between particles can not be excepted), for them invariants of physical values do not exist. It means, that LPS forget their initial conditions, therefore for them there is no solution at a level of trajectories (though at a probabilistic level the solution exists). According to above-mentioned, for investigation of self-organization mechanisms origin in complex systems within the framework of the present project the detail learning is necessary, on the one hand, both models CDS and models LPS, and on the other hand, also conditions and ways of their unification, reduction and addition.

Within the framework of the project the full technological sequence of receiving, processing, analysis of data and simulation of environments will be developed. Technical basis for realization of full-scale researches will be lidar complex earlier created by the Performer, which one will be also advanced pursuant to the project tasks. With the purpose of effective and reliable ecological monitoring (in particular, when the means of a laser detection appear incapable) in the project principles and ways of radiolocation (radar) measurement of environments parameters will be also developed. Besides, for the first time in practice of ecological monitoring the gravitational control of an environmental ecological condition with use of the data obtained on the basis of gravimeters will be carried out.

Expected Results and their Application. The obtained theoretical results in a course of project realization will be aimed on investigation of LPS models and problems of dissipative structure formation in thermodynamics, hydrodynamics and gravidynamics, and also solving recognition problems of chaotic signals of different physical nature (lidar, radar and gravimeteric). The obtained practical result within the framework of the project will be directed at creation of soft-hardware of an environmental monitoring for estimation and prediction of ecological situation on the one hand and incidentally will be used for diagnostic and detection of pathology of different people's diseases on the basis of developed means of speech and medical signals recognition on the other hand. In particular, within the framework of solving project problems the following results will be obtained:


• For the first time, the Boltzmann equation for the operators of density in configuration and momentum (impulse) spaces, which assumes the use of singular potentials of interfrequency interaction will be used for learning dynamic behavior of LPS models one- and multi-component gas media. The creation of discrete version of this equation will allow to estimate the additional information for definition of products of Schwartz's distributions and to develop effective algorithms for prediction of dynamic behavior of one- and multi-component gas media.

• The evolutional equations describing the process of slow flowing gravitational tightening of the spheroidal body based on LPS model, and also the equation of a weak gravitational field will be obtained. On the basis of these equations the capabilities of origin of evolutional structures in the gravitating spheroidal body will be researched. The statistical model of gravitation suggested in the project, has no analogs in the world science. However, besides only scientific sphere of application, measurable results and statistical data processing obtained during experiments with the use of gravimeter, will have priority meaning in the world practice of ecological situation monitoring.

• The method of definition of self-organization conditions in CDS will be offered and it will be shown, that by means of methods of Lyapunov stability theory it is possible not only to define the self-organization conditions but also to calculate bifurcation parameter values of researched systems with a high accuracy.

• The method of the multifractal analysis for investigation of multidimensional fractal sets (first of all, for attractors reconstructed from researched time series) will be developed and the computer simulation of diffusion - limited aggregation structure is carried out through algorithm of self-noncrossing random walks.

• The chaotic signal processing methods and algorithms will be developed. Including the effective algorithms of training of autoregressive hidden Markov models and the new approaches to training of neural networks having faster convergence, in particular, approaches founded on a linearization of neural network models and level-by-level training with use of Kalman's filter theory will be offered.

The scientific results obtained within the framework of the given project, will have two important areas of practical application. At first, they will be directed at creation of regional systems of the Earth's atmosphere monitoring, secondly, they will be directed at creation of hard-software for diagnostic and detection of a pathology of different people's diseases. In particular, will be developed:


• The data processing methods of a multiwave laser probing, which are optimal for detection of hidden legitimacies of transformation of a stratosphere aerosol layer.

• The monitoring methodics, founded on computer simulation of processes of pollution transportation, where the input data of model are determined on the basis of the experimental lidar and radar data.

• The analysis methods and algorithms of multifractal properties of biomedical signals with the purpose of reliable estimation of dynamics of organism's pathological processes and detection of hidden violations of the cardiovescular system regulation.

• Methods and algorithms of automatic diagnostics of such pathological diseases in speech formation human organs as one-side larynx paralysis, ginotonus dysphonia, hordit, nodules of phonatory bands.

Within the framework of the project new fundamental knowledge about methods of self-organization in complex systems will be obtained and the algorithms of the effective environmental monitoring based on lidar, radar and gravimeteric measurements will be developed.

Meeting ISTC Goals and Objectives. As a result of project implementation the new fundamental knowledge of methods of self-organization in complex systems will be obtained and the algorithms of effective environmental monitoring will be designed. On the present moment the main participants of the project in this area performed large preliminary work and the main directions of researches are determined. The project goals, its actuality and proposed approaches to the project realization were discussed and supported by the leading world-wide specialiststhe project collaborators and also the Nobel Prize winner, Professor I. Prigogine and Dr. I. Antoniou (Belgium/ USA).

The importance of present project is to provide a possibility for the scientists who worked at the military field (in particular, in the field of nuclear weapons delivery and means of its transportation) to deal with fundamental and applied research problems in the interests of industrial and commercial application of their activity results for welfare of modern society development.

This project will be the basis to unite research forces of the Belarusian scientists working in the field of self-organization systems modeling, digital signal processing and ecological monitoring with the aim of the common association forming within the framework of International Scientific Society, such as AMS, AIP, SPIE, IEEE, EURASIP, URSI and other.

The work within the framework of the project will positively influence on the Performer activity personally of two academic institutes and two leading institutions of higher education in the Republic of Belarus by means of material and moral incentive of the scientists and their students and post-graduates involved in the project, to create suitable conditions for their fruitful scientific and methodical activity based on modern computational and apparatus tools.

The present project, doubtless, will encourage federating the scientists, attracted in the project, and also their colleagues, young scientists and post-graduate students in world scientific community by means of their participation at international conferences, symposia and workshops. It will promote the solution of some technological and economical problems occurring now among the scientists in countries of former USSR, and also development, in the whole, international relations in the frameworks of «national diplomacy».

Duration of the project - 36 months, total person-months of project effort 342.45 man.*months.

The role of the project Collaborators. The cooperation with the project collaborators will promote: to fruitful exchange of the scientific information within the framework of project researches, impartial assessments and reviewing of scientific outcomes of the project, active participation in development of practical results, direct assistance and support for the project.

In particular, with Professor, Dr. Martin Vetterli (USA / Switzerland) it is expected fruitful and useful cooperation in the field of dynamic systems modeling, analysis and synthesizing of digital filters, new algorithms of information compression based on wavelet transforms, methods for recognition of signals from sensors of ecological monitoring systems.

With Professor, Dr. Maurice Kibler (France) the useful cooperation is planned in the field of non-linear methods for the analysis of complex systems, models of lattice gas, models of self-organization in the active media.

With Professor, Dr. Cihan H.Dagli (USA) it is expected cooperation in the field of methods of self-organization in the complex systems, analysis and synthesizing of artificial neural networks, designing of smart engineering systems for recognition of speech signals, genetic algorithms and hidden Markov models, identification of the data lidar, radar and gravimetric measurements.

With Dr. Kevin Priddy (USA) the fruitful cooperation is planned in the area of methods and means of nonlinear dynamics, algorithms and software of the fractal analysis, soft-hardware tools for digital signal processing in the ecological monitoring systems.

With Professor, Dr. Nikos Mastorakis (Greece) it is expected useful cooperation in the field of methods of dynamic systems modeling, including multidimensional, algorithms for digital filtering, methods of self-organization in the neural networks.

Technical Approach and Methodology. For computer modeling of physical structures in thermodynamics, hydrodynamics and gravidynamics based on models of lattice gas the methods of the cellular automata theory (i.e. neural networks consisting of elements, changing their states at the discrete moments of time under the dynamic or stochastic law depending on what was a state of the element and its neighbours on a network at the previous discrete moment of time) will be used and applied for new types of the dynamic equations. These methods allow to carry out direct imitation of dynamic processes with use in parallel working computer networks.

In study of processes of self-organization of CDS it is supposed to prove, that use of methods of the Lyapunov theory of stability allows not only to define of self-organization conditions, but also to calculate with high accuracy of meanings of bifurcation parameters of researched system without the numerical solution of the differential equations.

In study of processes of self-organization the method of the multifractal analysis also will be used. In particular, the algorithms of self-noncrossing random walks will be used for computer modeling of diffusion- limited aggregation.

In problems of diagnostics and the control of a state of CDS will be used models for calculating the characteristics of autowave processes.

For digital processing of signals produced by CDS of a various physical nature, the methods of the local - topological analysis allowing completely to determine topological structure of attractors and scale invariancy properties of investigated time series will be used.

The methodology of prediction of chaotic signals as time series will use the generalized hidden Markov models (in particular, autoregression hidden Markov models) and genetic algorithms.

The methodology of laser sounding will be developed on the basis of increasing information content of lidar systems as a result of extending their spectral range and designing new algorithms for digital data processing with use of the CDS analysis methods. Experimental lidar investigations of aerosol pollution and atmospheric stratospheric layer will be carried out by utilizing the lidar complex comprising multifrequency stationary and mobile lidar devices operating over ultraviolet, visible and near infrared spectral ranges.

The statistical methods of synthesis and analysis of precision algorithms of automatic measurement of parameters of an environment and adaptive recognition of meteorological formations based on correlated polarized and multifrequency radar-tracking portraits will be applied to radar-tracking monitoring of environmental processes.

In the methodology area of gravitational monitoring of ecological conditions the project realization result will be statistical model and statistical methods of data processing of gravitational measurements and monitoring techniques.


Back