Robust Diagnosis of Dynamic Systems
Funding: Grant Agency of Ministry of Education and Academy of Science
of Slovak Republic VEGA, Grant No. 1/6270/99
Duration: 1999 - 2002
Research activity of the group is concentrated on:
- investigation of test signals which reflects incosistencies between the
fault-free and fault system as well as neural network classifier training
and discriminative hidden Markov models learning for system fault classes
- testing of procedures for fault function decomposition and fault isolation
- robustification of fault tolerant control for uncertain dynamic systems
Scientific goals/research targets:
- verification of application-oriented structual modelling and model-based
control and diagnosis algorithms for uncertain dynamic systems
- development of methods and algoritms for maximum likelihood risk-sensitive
estimation of residuals in robust diagnosis systems based on analytical
redundancy
- combining model-based residual generation, neural decision making and
stochastic modelling of the system fault states to give the appropriate
procedures
In the year 1999 the research effort was spent on the analyses of robust
residual generation procedures, risk-sensitive residual estimation in robust
diagnosis as well as on methods to handle uncertainty in models of dynamic
systems, stochastic models of faults and neuro-based decision making.