Design of gratings and frequency selective surfaces using Fuzzy ARTMAP neural networks

Abstract

This paper presents a study of the Fuzzy ARTMAP neural network in designing cascaded gratings and frequency selective surfaces (FSS) in general. Conventionally, trial and error procedures are used until an FSS matches the design criteria. One way of avoiding this laborious process is to use neural networks (NNs). A neural network can be trained to predict the dimensions of the elements comprising the FSS structure, their distance of separation, and their shape required to produce the desired frequency response. In the past, the multi-layer perception architecture trained with the back-prop learning algorithm (back-prop network) was used to solve this problem. Unfortunately, the backprop network experiences, at times, convergence problems and these problems become amplified as the size of the training set increases. In this work, the Fuzzy ARTMAP neural network is used to address the FSS design problem. The Fuzzy ARTMAP neural network converges much faster than the back-prop network, and most importantly its convergence to a solution is guaranteed. Several results (frequency responses) from cascaded gratings corresponding to various angles of wave incidence, layer separation, width strips, and interstrip separation are presented and discussed.

Distributed Learning, Recognition, and Prediction by ART and ARTMAP Neural Networks

“…ART neural network models have added a series of new principles to the original theory and have realized these principles as quantitative systems that can be applied to problems of category learning, recognition, and prediction. Applications of unsupervised ART networks (Carpenter & Grossberg, 1987Carpenter, Grossberg, & Rosen, 1991) and supervised ARTMAP networks (Carpenter, Grossberg, Markuzon, Reynolds, & Rosen, 1992;Carpenter, Grossberg, & Reynolds, 1991) include a Boeing parts design retrieval system (Caudell, Smith, Escobedo, & Anderson, 1994), satellite remote sensing (Baraldi & Parmiggiani, 1995;Gopal, Sklarew, & Lambin, 1994), robot sensory-motor control (Bachelder, Waxman, & Seibert, 1993;Baloch & Waxman, 1991;Dubrawski & Crowley, 1994;Srinivasa & Sharma, 1996), robot navigation , machine vision (Caudell & Healy, 1994), 3D object recognition (Seibert & Waxman, 1992), face recognition (Seibert & Waxman, 1993), automatic target recognition (Bernardon and Carrick, 1995;Koch, Moya, Hostetler, & Fogler, 1995;Waxman et a!., 1995), medical imaging (Soliz & Donohoe, 1996), electrocardiogram wave recognition (Ham & Han, 1996;Suzuki, 1995), prediction of protein secondary structure (Mehta, Vij, & Rabelo, 1993), strength prediction for concrete mixes (Kasperkiewicz, Racz, & Dubrawski, 1994), signature verification (Murshed, Bortozzi, & Sabourin, 1995), tool failure monitoring (Ly & Choi, 1994;Tarng, Li, & Chen, 1994;Tse & Wang, 1996), chemical analysis from UV and IR spectra (Wienke, 1994), digital circuit design (Kalkunte, Kumar, & Patnaik, 1992), frequency selective surface design for electromagnetic system devices (Christodoulou, Huang, Georgiopoulos,...…”

Tecnología aplicada en el funcionamiento y la detección de minas antipersonales: estado del arte

“…Por un lado, se aplican algunos sistemas de técnicas gráficas tomografía [9,10], así como aplicaciones usando ecuaciones de las integrales [11] , pero estos han tenido un éxito solo parcial, debido principalmente a la complejidad de los datos de campo, que contiene altos niveles de ruido causados por no homogeneizar los medios de acogida. Por otro lado, las técnicas basadas en redes neuronales (NN) que aprenden a adaptarse basándose en las experiencias recogidas en diferentes topologías [11][12][13][14][15][16][17] conocidas como neural network, se han propuesto para resolver el pr o blema del electromagnetismo y la inversión canónic a . Por ejemplo, un esferoide incrustado en un medio de acogida [17,18], y otras mejoras en relación con las formas geométricas más realistas para aplicaciones ingenieriles [18,19], aun incluyendo la consideración de un medio de acogida no homogénea [19,20].…”

Particle-Swarm-Optimization-Based Selective Neural Network Ensemble and Its Application to Modeling Resonant Frequency of Microstrip Antenna

“…Moreover, the computation of the mapping is convenient and fast. Because NN is very suitable for modeling and optimization of complex electromagnetic systems that face CAD optimized process, it is widely used in electromagnetic field [14] [15] .…”