Microwave seminar | 30 October 2025

This paper presents a rapid machine learning-based method for optimizing reflector antenna feeds. Calculating the aperture efficiency of reflector antennas with a given feed using numerical simulation is computationally intensive. We propose a simple regression model, trained on numerical simulation data, that predicts reflector antenna aperture efficiency from seven features extracted from the feed's radiation pattern. A specific post-processing routine in CST Studio Suite automatically calculates these features from the feed design and estimates aperture efficiency with 10% accuracy, bypassing the need for full reflector antenna numerical simulation. The proposed approach drastically reduces design time and facilitates efficient feed geometry optimization.

Main paper/arXiv, related to the seminar, and other references

Сивов Н. О., Лемберг К. В. Оптимизация облучателя с ближнепольным контррефлектором с использованием алгоритмов машинного обучения // Материалы VIII Всерос. науч.-техн. конф. «Системы связи и радионавигации». 2024. С. 110-115.

Presentation will be held in Russian

Real-time cell concentration monitoring instruments for cell growth in bioreactors have shown an evident increase in production at industrial levels, however, the cost of existing technologies hinders its implementation in small and middle sized companies in developing countries. I will demonstrate the developing process for a low-cost instrument prototype that implements polymeric materials compatible with cell cultures sterilization; conditioning electronic systems for generating and receiving light intensity signals, and noise-treatment configurations to mitigate temperature and electromagnetic effects. The final low-cost product complies with industrial regulations and achieves real-time determination of cell concentration by its correlation with its optical density.

Presentation will be held in English