JÍLEK, L. Automatizovaný návrh planárních antén v jazyce Python [online]. Brno: Vysoké učení technické v Brně. Fakulta elektrotechniky a komunikačních technologií. 2025.
The project was aimed to develop an artificial antenna designer in Python. Models were to be created using PyAEDT library, simulations results were to be used to train a classifier, and selected antennas were to be implemented and measured to calibrate numerical models. For users, a guide was to be written. Using PyAEDT, scripts for modeling and optimization of patch antennas, planar monopoles, substrate integrated horns and planar dipoles were written and explained. Models were used to train a neural classifier for selecting an optimal antenna. For the code, user interface was created using TKinter library. To calibrate models, a wideband monopole and a patch antenna were manufactured and measured. Obviously, objectives of the project were met. Nevertheless, more time and energy had to be devoted to writing the report. The descriptions are very brief and incomplete (only the patch antenna script is presented, user manual is very brief, description of classifier and GUI implementation are missing, diagram of the code is not provided, simulations and measurements are not in the same chart to allow simple comparison, etc.). Student worked on the project independently, and regularly discussed substantial problems. A deeper knowledge of antennas was expected but programming abilities were good. A relevant amount of work on the project was done.
The thesis is focused on creation of an application that would automate the design and optimization of planar antennas, based on the required parameters. In the first section, the student explains the motivation behind the activity and briefly familiarizes the reader with the tools that were used to develop the application. Given that the thesis is built mainly on the PyAED library, I would expect a more detailed description of this part. Further in the section, the student describes the types of the antennas that were considered. Here I miss larger discussion and comparison of the various concepts, including their disadvantages and design limitations, that might help to further improve the selection criteria used by the tool to determine the suitable type of the antenna for intended usage, ensure that the optimization part will converge and reduce the time needed to carry out the task. In some cases, it is not obvious, from which reference the design equations were taken and also, some of the equations were incorrectly re-written, such as equation (2) that calculates the effective permittivity of a patch antenna. The second section is focused on practical implementation of the application. The student derives initial dimensions of the antennas, based on the defined formulas, and creates the appropriate numerical model in HFSS. Next, the optimization part is discussed where I miss a more detailed comparison of the considered optimization methods, ideally including concrete data supporting the final method selection. By running simulations, base parameter data sets across all the antenna types were obtained that, combined with user input and preferences, are used to select the most suitable type of the antenna that will later be optimized. I find this approach as suitable and logical. To improve user experience, a GUI was created however, the concept is very outdated. Instead of Tkinter, the QT library could have been selected, which is not much more complicated for use and allows to create modern interfaces. Functionality of the resulting application was demonstrated by running a series of design tasks for various antenna types and target center frequencies. Some of the antennas were manufactured where the simulated and measured properties were compared. Presentation of the results is not ideal, as the simulated and measured results are not displayed in one graph for better comparison (Figure 7 and Figure 8 for example) or lack the simulated results completely (Figure 11). This makes it difficult to verify how accurate the generated models are in terms of real performance however, from the comparison of the S11 characteristics, the simulated and measured results are in good agreement across the antenna type portfolio. The discrepancies found in the results are discussed and reasons identified. In overall, the objectives of the thesis were met and the student delivered an application that offers the required functionalities. The properties of the resulting designs match with the theoretical assumptions. As for the reproaches mentioned mainly towards the formal aspects of the thesis, those are not seen as critical, though they reduce the overall impression to some extent. I recommend the thesis for defense.
eVSKP id 167873