1.

Product quality assessment by means of approximate reasoning

The application of approximate reasoning significantly decrease the detection instability, i.e. frequent switchings between alarms. The approach can be viewed as a complement to the statistical tests. An implementation on a prototype for quality monitoring of electrically commutated motors is succesfully done.

F.07 Improvements to an existing product

COBISS.SI-ID: 24699431

2.

Parameter estimation of nonlinear dynamic systems

In case of non-stationary production condition to describe the processes one needs non-linear dynamic models. The dissertation comes with new ideas on estimation of the model parameters in the stochastic setup by relying on various approximations of the probability density functions.

D.09 Tutoring for postgraduate students

COBISS.SI-ID: 256638720

3.

Production control by means of neural models

Moder control of manufacturing systems acquire a huge amount of process data. The problem is how to extract informtation out of the data in order to provide useful support to the operators in achieving the required key performance indicators. The approach utilizing neural models has been elaborated and demonstrated on a complex simulated process.

B.06 Other

COBISS.SI-ID: 24594215