Robust real-time determination of dielectric permittivity from reflection data using a neural network
DOI:
https://doi.org/10.15421/332518Keywords:
dielectric permittivity, reflection coefficient, neural networks, CNN, non-destructive testingAbstract
This paper presents a neural network approach for reconstructing dielectric permittivity in the 38 –52 GHz band. By employing logarithmically compressed time-domain features of the inverse reflection coefficient, the developed convolutional neural network (CNN) model achieves a relative error of 2 – 3 %. The method enables accurate, non-iterative material characterization suitable for rapid analysis.
