Filters
cris logo
-
New search Filters
Select from list
Displayed from Show more
You have reached the maximum number of displayed search results.
All results displayed
No results are available for the search performed
Displayed from
You have reached the maximum number of displayed search results.
All results displayed
Loading results
Obtaining statistical data

Projects / Programmes source: ARIS

Identifikacija in analiza modelov za načrtovanje vodenja dinamičnih sistemov na podlagi Gaussovih procesov (Slovene)

Edit
Research activity

Code Science Field Subfield
2.06.02  Engineering sciences and technologies  Systems and cybernetics  System theory and control systems 

Code Science Field
2.02  Engineering and Technology  Electrical engineering, Electronic engineering, Information engineering 
Evaluation (rules)
source: COBISS
Evaluation of bibliographic research performance indicators according to ARIS methodology
Citations Citations for bibliographic records in COBIB.SI that are linked to records in citation databases
source: WoS source: Scopus source: COBISS
Researchers (4)
no. Code Name and surname Research area Role Period No. of publications
1.  16161  PhD Samo Gerkšič  Systems and cybernetics  Researcher  2009 - 2012  136 
2.  05807  PhD Nadja Hvala  Systems and cybernetics  Researcher  2009 - 2012  208 
3.  10598  PhD Juš Kocijan  Systems and cybernetics  Head  2009 - 2012  450 
4.  24269  PhD Bojan Musizza  Energy engineering  Researcher  2009 - 2012  117 
Organisations (1)
no. Code Research organisation City Registration number No. of publications
1.  0106  Jožef Stefan Institute  Ljubljana  5051606000  90,768 
Significance for science
Until recently, a reliable data driven method that accounts for uncertainty in a natural manner has been lacking. The main innovation of this project is remedying this lack by the further development of a practical framework for modelling nonlinear dynamic systems from uncertain information with purpose to be used in control design. Due to the rapid increase in computing power over the last decade, probabilistic approaches to modelling have attracted increasing interest. However, the statistics literature, where much of these developments are discussed, is largely concerned with static systems. We proposed to use the Gaussian process modelling methodology for dynamic systems. The proposed approach employs an entirely probabilistic description of the system using Gaussian processes in a Bayesian context whereby a prior is conditioned on the available information. The adaptation of this methodology to the modelling of dynamic systems in an engineering context and application of these models for control design open up new possibilities for data driven modelling and safe and robust control design. Being nonparametric, the modelling is attractive to engineers. Research considered also dealing with two other related problems beside development of new methods and upgrade of already existing research on methods for experimental modelling, i.e., identification, of dynamic nonlinear systems with Gaussian process models. The first one is decrease of computational burden for model learning at identification which was dealt with multicore-processor hardware and novel algorithms. Successful results are potential advancement in adopting developed modelling methods in engineering practice. The second problem is the analysis of nonparametric models for their use in control systems design. Results in this case will also increase applicability of Gaussian process models that are already very attractive due to the content of information they possess. The Gaussian process modelling methodology and control design were tailored to dynamic systems. However, the results can be used for static systems and can be readily extended for other applications beside control design. A library of software functions embodying the developed techniques was developed and its utility was demonstrated by numerous applications. The main contributions are as follows: - Directions for selection of covariance functions, that are the basic elements for modelling with Gaussian process models. - Systematic presentation of model structures based on used regressors and validation of models' properties. - The research of online/recursive identification of Gaussian process models, - The method for identification of model with fixed structure and varying parameters, where varying parameters are modelled with Gaussian process models. - The method for the identification of Wiener and Hammerstein Gaussian process models. - An algorithm for the accelerated learning of Gaussian process models with parallel computation on graphic process unit. - An algorithm for online/recursive identification of Gaussian process models, - Applications of developed methods on various environmental, urban traffic and process engineering dynamic systems. We estimate the contributions as a step forward in the research of nonlinear systems identification.
Significance for the country
Several applications were used to demonstrate the utility of the Gaussian process modelling methodology and control design, specifically, its application to chemical process plant, wastewater treatment plant, urban transportation system, rotational machinery, etc. One of the applications in this project was modelling of the gas-liquid separation plant that has been used before to demonstrate utility of various control algorithms for industrial practice. A further application is modelling and control of different sorts of wastewater treatment plants that are complex systems that represents biotechnological plants. A probabilistic framework capable of representing complex systems is indicated. The proposed framework employs models that provide a complete probabilistic description. Modelling of queue lengths in urban traffic, modelling of ozone pollution and a system for prognosis of gear health in rotational machinery were further used to demonstrate the utility of Gaussian process models for modelling dynamic systems. The relevance of the proposed modelling framework and control design to selected process and bio-technological application is indicative of its relevance to complex engineering systems, which can be frequently met also in other fields of industrial production. Their control design and efficiency depends on the use of models. Hence, a general modelling framework and corresponding control design methods has a potential for great strategic value to the engineering community. Relevance of the proposed project to the development in Slovenia can be summarised as follows. • The project is proposing a new approach for modelling and control design of complex dynamic systems from uncertain data and therefore has all the potential to facilitate development in some fields important for Slovenia. • Fields where proposed modelling and simulation framework can be applied is, as demonstrated, wastewater treatment plant technology and environmental science, i.e., air pollution modelling and forecasting. • The other field that may be affected is automation in the field of manufacturing, i.e., rotational machinery, and process engineering, i.e. gas-liquid separators. Consequences are improved quality of industrial production. A channel for the dissemination of the results of the project to a wider audience beyond the research and engineering community was the educational process. Researchers in the project are involved in University education as instructors and they taught and promoted methods for efficient modelling, simulation and control of complex dynamic systems from uncertain data and consequently gave a new quality to the content of Slovene education.
Most important scientific results Annual report 2009, 2010, 2011, final report, complete report on dLib.si
Most important socioeconomically and culturally relevant results Annual report 2009, 2010, 2011, final report, complete report on dLib.si
Confirmation required
Views history
Favourite