Search academic texts

Information × Registration Number 0226U002838, (0124U000966) , R & D reports Title Detection and classification of explosive objects on the surface and inside the ground by ultra-wideband pulse georadar and neural networks popup.stage_title Комп’ютерне моделювання перетворень полів на поверхні ґрунту і розпізнавання об’єктів Head Plakhtii Vadym A., Доктор філософії Registration Date 07-03-2026 Organization V.N. Karazin Kharkiv National University popup.description1 The goal of the project is to create a system for the automatic detection and recognition of explosive devices and objects, including those that have no metal parts, using impulse ultrawideband radar and a unit for processing received pulses based on artificial neural networks. To achieve the project goal, it is proposed to synergistically combine the analytical approach of the Tretyakov evolutionary equation method and the numerical calculation of fields by the FDTD method for a comprehensive analysis of field transformations, existing statistical methods for processing non-stationary signals and a deep artificial neural network to solve the problem of recognizing dangerous hidden objects. At the same time, it is planned to perform on this basis the task of calculating the characteristics of the positioning system on impulse ultrawideband fields. popup.description2 During the reporting stage, a set of studies was carried out aimed at investigating the propagation of transient electromagnetic waves in near-surface media and exploring the use of impulse signals for subsurface object detection and spatial positioning tasks. A theoretical analysis of analytical solutions describing the radiation and propagation of transient electromagnetic waves was performed, allowing the key features of the spatial–temporal structure of electromagnetic fields near media interfaces to be identified. Numerical modeling was conducted to investigate the interaction of impulse electromagnetic signals with soil as well as with dielectric and metallic inclusions typical for subsurface sensing problems. As a result, time-domain dependencies of electromagnetic fields at different spatial points were obtained, making it possible to identify characteristic signal features associated with the presence of hidden objects. The obtained time signatures were used to form datasets suitable for machine learning applications. The potential of artificial neural networks for automatic recognition of subsurface object signals and estimation of their parameters was analyzed. Special attention was also paid to the study of impulse electromagnetic radiators and the analysis of their time-domain radiation patterns. It was demonstrated that the waveform of the received signal changes depending on the observation angle, which enables the use of these characteristics in spatial positioning systems. The obtained results can be applied to the development of new approaches to ground-penetrating radar signal processing, improvement of subsurface sensing systems, and creation of local positioning technologies based on ultra-wideband impulse signals. Product Description popup.authors Nadiia P. Yelisieieva Oleksandr M. Dumin Hennadii P. Pochanin Svitlana V. Pshenychna Serhii L. Berdnyk Oleksandra S. Skvortsova popup.nrat_date 2026-03-07 Close