Tayfun Günel

Dielectric Permittivity Estimation of Cylindrical Objects Using Genetic Algorithm

A new method based on a genetic algorithm for the estimation of the dielectric permittivity distribution of a cross section of an inhomogeneous cylindrical object illuminated by a low power microwave radiation field is presented. This method has been tested for some objects using simulated scattered fields. A good convergence to the expected values was obtained.

Application of Continuous Parameter Genetic Algorithm to the Problem of Synthesizing Bandpass Distributed Amplifiers

Summary Continuous parameter genetic algorithm (CPGA) is applied tosynthesis of monolithic GaAs FET bandpass distributed amplifiers for desired flat gain and maximum cutoff bandwidth. The results obtained by CPGA are compared with the results obtained by fuzzy genetic algorithm (FGA) and binary genetic algorithm (BGA).

NUMERICAL MODELING OF MICROSTRIP RADIAL STUB

A simple numerical algorithm used to model microstrip radial stub is presented. The cascaded interconnections of uniform transmission lines with incremental length dl are considered for the calculation of input impedance of a radial stub. Calculated results are consistent with the theoretical and experimental results found in the literature.

Hybrid optimization approach to the pattern synthesis of conformal antenna array

In this paper, Hybrid Taguchi Genetic Algorithm (HTGA) is used to minimize the side lobe level and to keep half power beam width in certain values for cylindrical conformal antenna array. The HTGA is based on Genetic Algorithm (GA) and Taguchi Algorithm (TA). Results obtained by GA, TA and HTGA are compared. The best results are also compared with the results obtained by HFSS (High Frequency Structural Simulator).

Conformal antenna array synthesis using Taguchi Algorithm

In this paper, Taguchi Algorithm is used to design a cylindrical conformal antenna array with desired side lobe level and half power beam width. Results obtained by traditional genetic algorithm and Taguchi algorithm are compared. The desired side lobe level and half power beam width are successfully obtained by Taguchi algorithm more faster than genetic algorithm.

Synthesis of Radial Stubs Using a Numerical Model

An optimization approach to the synthesis of radial stubs using numerical model is presented. Microstrip and coplanar waveguide radial stubs are considered. The results are compared in terms of the lengths of the radial stubs.

A New Hybrid Approach to Radar Target Classification for the Estimation of Scattering Centers

Radar images, range profiles and scattering centers are used as feature parameters in radar target classification applications. Scattering center parameters, when used as feature parameters, enable an efficient compression of feature space compared to classical target classification methods based on radar images and range profiles. A method used for the estimation of scattering centers via cancellation of side lobes is the CLEAN algorithm. In this work, model based Prony, MUSIC, ESPRIT and evolutionary based CLEAN methods are applied for the estimation of scattering centers. A hybrid method is proposed which improves the convergence of evolutionary based CLEAN. Scattering centers which are estimated by aforementioned methods are classified using correlation based matching score method, Bayes classifier and artificial neural networks. Classification is accomplished using simulated data of four different aircraft models created by the point target model at different frequency bands and aspect angles.

A hybrid method based on genetic algorithm approach to the estimation of scattering centers for radar imaging

In this work, a novel scattering center extraction method using genetic algorithm based CLEAN is proposed. Initial values for optimization parameters are estimated from the power spectra of the measured data using the Fourier transform. The proposed method reduces computational time compared to existing evolutionary programming (EP) based approaches.

Optimization on wedge type absorbers via genetic algorithm

This work presents the optimal design parameters of a V-shaped dielectric absorber. The main parameters which affect the reflection coefficient are the dielectric permittivity, the length and the shape of the material. The main goal is to maintain the reflection coefficient minimum over a wide frequency band. For this reason, genetic algorithm is applied to determine the length of the absorber between 1-10 GHz. An investigation for the different dielectric permittivities is also carried out. The results give a convenient selection possibility for such a material for anechoic chambers.

A new approach to the optimal choice of a wavelet for signal representation

A new optimization approach for the design of discrete orthogonal wavelets of support less than or equal to some given integer that leads to the best approximation to a given finite support signal up to a desired scale is discussed. A new global hybrid optimization strategy, involving the combination of simulated annealing and Hooke-Jeeves algorithms is developed and applied to minimize certain cost functions defined by an error function between original and approximated signals and wavelet admissibility conditions. The method converts the constrained minimization over the parameters that define discrete finite support orthogonal wavelets into an unconstrained one and can be implemented much faster than any other algorithms presented in literature.