Title :
Enhanced neural modeling of planar antennas and filters
Author :
Petr Smid;Zbynik Raida;Jioi Horak
Author_Institution :
Brno University of Technology, Dept. of Radio Electronics, Purky?ova 118, 612 00, Czechia, Czech Republic
Abstract :
The paper deals with an enhanced neural modeling algorithm based on incorporation of both, global optimization method and local one. Proposed algorithm depresses disadvantages of classical neural modeling. A neural model of the low-pass filter is trained to behave as a numerical one, which is analyzed in Zeland IE3D. Both, the neural model and the numerical one are optimized using particle swarm optimization method (PSO). Finally, the results of optimization are compared.
Keywords :
"Planar arrays","Filters","Impedance","Artificial neural networks","Numerical models","Optimization methods","Cutoff frequency","Neural networks","Feedforward neural networks","Feedforward systems"
Conference_Titel :
Antennas and Propagation, 2006. EuCAP 2006. First European Conference on
Print_ISBN :
978-92-9092-937-6
DOI :
10.1109/EUCAP.2006.4584903
