Реєстраційний номер
2119U010511, Матеріали видань та локальних репозитаріїв
Категорія
Стаття
Назва роботи
Methods and means of processing discrete information in networks with a high level of noise
Автор
Kvetnyy R. N.
Kulyk Y. A.
Knysh B. P.
Garmash V. V.
Romaniuk R. S.
Sagymbekova A.
Кветний Р. Н.
Кулик Я. А.
Книш Б. П.
Гармаш В. В.
Романюк Р. С.
Сагимбекова А.
Дата публікації
16-01-2019
Постачальник інформації
Вінницький національний технічний університет
Першоджерело
http://ir.lib.vntu.edu.ua/handle/123456789/23231
Видання
Society of Photo-Optical Instrumentation Engineers
Опис
The article is devoted to increasing the efficiency of digital signal processing in the conditions of high level of interference, for which the efficiency and reliability of the transmission of information have a priority over the speed of transmission and the amount of CP resources used. The authors provide readers with the improvement of current methods in order to increase the performance of information transmitting in difficult conditions environment. The method of determining the decomposition coefficients, which uses the replacement of the biorthogonal coefficients of the wavelet decomposition with the approximation sum using a series of quasi-random delta sequences, is used in the work, which is used to eliminate the Gibbs effect in signal processing. The method for evaluating the spectrum of the signal for an adaptive threshold method, which uses a multi-window average estimation of the logarithmic spectrum of the signal, is improved. A method of the fast median filtration which processes the finite quantities of date vector with splitting an original data vector onto some parts has been developed. The method of parallel fast wavelet transform is improved, which uses the partition of the data vector into blocks for processing data using a local wavelet transform in the diagonal sequence. The theoretical researches and modeling demonstrate the significant efficiency of the newly proposed and improved methods.
Додано/оновлено в НРАТ
2026-10-29
