BAYESIAN DECISION MAKING TO SUPPORT CHANGE DETECTION IN COMPLEX MANUFACTURING SYSTEMS
Project goals
An enduring problem in FDI is that of characterizing the dynamics of complex systems. This is a key step towards feature extraction. The difficulty stems from the fact that in many cases it is not possible to derive a suitable model from the first principles. Given historical data, the typical way to proceed is via data projection into latent structures or by using integral transforms (orthogonal decomposition). Two main topics are addressed within the proposal: (1) Employment of Bayesian framework as systematic approach to feature extraction via exploiting on-line and historical process data available. The considered solution supposes dynamic probabilistic mixture modelling of the underlying process. The probabilistic extraction of information from process data allows us to overcome the dependence on application and does not require detailed knowledge of system dynamics. Furthermore, the related algorithms and used techniques mostly have a long tradition and already proved to be efficient and acc
Keywords
Modelling of complex systemsFault detectionBayesian estimationAdaptive control
Public support
Provider
Ministry of Education, Youth and Sports
Programme
KONTAKT
Call for proposals
KONTAKT 6 (SMSM2008ME3)
Main participants
—
Contest type
VS - Public tender
Contract ID
9186/2008-32
Alternative language
Project name in Czech
BAYESIAN DECISION MAKING TO SUPPORT CHANGE DETECTION IN COMPLEX MANUFACTURING SYSTEMS
Annotation in Czech
An enduring problem in FDI is that of characterizing the dynamics of complex systems. This is a key step towards feature extraction. The difficulty stems from the fact that in many cases it is not possible to derive a suitable model from the first principles. Given historical data, the typical way to proceed is via data projection into latent structures or by using integral transforms (orthogonal decomposition). Two main topics are addressed within the proposal: (1) Employment of Bayesian framework as systematic approach to feature extraction via exploiting on-line and historical process data available. The considered solution supposes dynamic probabilistic mixture modelling of the underlying process. The probabilistic extraction of information from process data allows us to overcome the dependence on application and does not require detailed knowledge of system dynamics. Furthermore, the related algorithms and used techniques mostly have a long tradition and already proved to be efficient and acc
Scientific branches
R&D category
VV - Exeperimental development
CEP classification - main branch
BD - Information theory
CEP - secondary branch
BC - Theory and management systems
CEP - another secondary branch
JS - Reliability and quality management, industrial testing
10102 - Applied mathematics
10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)
20306 - Audio engineering, reliability analysis
Completed project evaluation
Provider evaluation
U - Uspěl podle zadání (s publikovanými či patentovanými výsledky atd.)
Project results evaluation
The project yielded an exchange of experience on decision methods and solved state filtering with unknown model matrices, developed partial forgetting for the case of different variability of individual parameters.
Solution timeline
Realization period - beginning
Jan 1, 2008
Realization period - end
Dec 31, 2008
Project status
U - Finished project
Latest support payment
Oct 29, 2008
Data delivery to CEP
Confidentiality
S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů
Data delivery code
CEP09-MSM-ME-U/02:2
Data delivery date
May 7, 2010
Finance
Total approved costs
80 thou. CZK
Public financial support
80 thou. CZK
Other public sources
0 thou. CZK
Non public and foreign sources
0 thou. CZK
Basic information
Recognised costs
80 CZK thou.
Public support
80 CZK thou.
100%
Provider
Ministry of Education, Youth and Sports
CEP
BD - Information theory
Solution period
01. 01. 2008 - 31. 12. 2008