prof. Ing. Václav Píštěk, DrSc.

1. Artificial intelligence methods for data simulation and identification in the area of ventilation

   Artificial intelligence methods belong among modern methods for system simulation and identification. These methods are especially suitable to apply to problems that are very difficult to solve by using classical mathematical methods or where the application of deterministic computation methods would require unacceptable simplification of the problem. Data identification by means of modern methods of artificial intelligence is an important task for solving problems of automatic data recognition in the area of ventilation. Approaches used by a graduant while solving the given task are as follows: Analysis of image recognition methods, analysis of neural networks and genetic algorithms for data identification.

   Tutor: Šťastný Jiří, prof. RNDr. Ing., CSc.

2. Solutions of continuous-time systems by the discrete methods

   Discretization is used for analysis and synthesis the continuous-time control systems. But is possible their use for solutions of continuous-time systems and for simulation continuous-time systems. It is new aspect of the analysis the continuous-time control systems. One of the simplest ways of discretizing or approximating a continuous-time plant is numerical approximation of differential equations. Difference equations can be obtained by discretizing differential equations. The other way of discretization is discretization by z transformation of transfer function G(s). In this suggest work is solution by Euler's method and bilinear method of this transformation and a method when the derivative term is simply replaced by a first-order difference expression and the integral by a sum. A new part of work will be solutions of unit step response and impulse response of continuous-time systems by discrete methods. It will contain an algorithm for their calculation.

   Tutor: Švarc Ivan, doc. Ing., CSc.