Pregled bibliografske jedinice broj: 483182
Determination of Complex Permittivity by Non-Destructive Method Using Neural Networks Modeling
Determination of Complex Permittivity by Non-Destructive Method Using Neural Networks Modeling // ICECom 2010 20th International Conference on Applied Electromagnetics and Communications Conference Proceedings / Bonefačić, Davor ; Bosiljevac, Marko (ur.).
Zagreb: Hrvatsko društvo za komunikacije, računarstvo, elektroniku, mjerenja I automatiku (KoREMA), 2010. str. 18-21 (predavanje, međunarodna recenzija, cjeloviti rad (in extenso), znanstveni)
CROSBI ID: 483182 Za ispravke kontaktirajte CROSBI podršku putem web obrasca
Naslov
Determination of Complex Permittivity by Non-Destructive Method Using Neural Networks Modeling
Autori
Vilović, Ivan ; Burum, Nikša ; Robert Nađ
Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni
Izvornik
ICECom 2010 20th International Conference on Applied Electromagnetics and Communications Conference Proceedings
/ Bonefačić, Davor ; Bosiljevac, Marko - Zagreb : Hrvatsko društvo za komunikacije, računarstvo, elektroniku, mjerenja I automatiku (KoREMA), 2010, 18-21
ISBN
978-953-6037-58-2
Skup
ICECom 2010 20th International Conference on Applied Electromagnetics and Communications
Mjesto i datum
Dubrovnik, Hrvatska, 20.09.2010. - 23.09.2010
Vrsta sudjelovanja
Predavanje
Vrsta recenzije
Međunarodna recenzija
Ključne riječi
Measurement; Dielectric permittivity; Neural network; Objective function; Reflection coefficient
Sažetak
A simple non-destructive method for determination of dielectric properties of arbitrary wall in building is presented. The free space method is based on the measurements of reflection coefficients for different angles of incidence. The walls constructed of unknown materials are modeled as homogenous dielectric plate with smooth surface. The measurements are made for the frequency of 2.4 GHz so the results can be applied to wireless local area networks (WLAN) planning and design. Complex permittivity is reconstructed using a neural network model trained by the calculated reflection coefficients for assumed permittivity. The inputs to the network are reflection coefficients and incident angles, while outputs are real and imaginary parts of the complex dielectric constant. The obtained results are compared with results obtained by a nonlinear-least-squares method.
Izvorni jezik
Engleski
Znanstvena područja
Elektrotehnika
POVEZANOST RADA
Projekti:
036-0361566-1570 - Elektromagnetski učinci i strukture u komunikacijskim sustavima (Šipuš, Zvonimir, MZO ) ( CroRIS)
275-0361566-3136
275-0000000-3260 - Integralna kvaliteta usluge komunikacijskih i informacijskih sustava (Lipovac, Vladimir, MZO ) ( CroRIS)
Ustanove:
Fakultet elektrotehnike i računarstva, Zagreb,
Sveučilište u Dubrovniku
