Home > All Issues > PIER
ISSN: 1070-4698

Vol. 81

Front:[PDF file] Back:[PDF file]
Latest Volume
All Volumes
All Issues

2008-04-24 PIER Vol. 81, 447-459, 2008. doi:10.2528/PIER08032101

Application of SVD Noise-Reduction Technique to PCA Based Radar Target Recognition

Kun-Chou Lee, Jhih-Sian Ou, and Ming-Chung Fang

The noise effect is very challenging in radar target recognition. It usually degrades the accuracy of target recognition and then makes the recognition unreliable. In this study, we present a noise-reduction technique to improve the accuracy of radar target recognition. Our noise-reduction technique is based on the SVD (singular value decomposition). The PCA (principal components analysis) based radar recognition algorithm is utilized to verify our noise-reduction scheme. In our treatment, the received signals are arranged into a Hankel-form matrix. This Hankel-form matrix is decomposed into two subspaces, i.e., the noise-related subspace and clean-signal subspace. The noise reduction is obtained by suppressing the noise-related subspace and retaining the clean-signal space only. Simulation results show that the accuracy of target recognition is greatly improved as the received signals are first processed by the SVD noise-reduction technique. With the use of proposed noise-reduction scheme, the radar target recognition can tolerate more noises and then becomes more reliable. The noise-reduction technique in this study can also be applied to many other problems in radar engineering.

2008-02-10 PIER Vol. 81, 437-446, 2008. doi:10.2528/PIER08011002

Directive Emission Obtained by Coordinate Transformation

Jing Jing Zhang, Yingliang Luo, Sheng Xi, Hongsheng Chen, Li-Xin Ran, Bae-Ian Wu, and Jin Kong

We use coordinate transformation theory to realize substrates that can modify the emission of an embedded source. Simulation results show that with proper transformation functions the energy radiated by a source embedded in these space variant media will be concentrated in a narrow beam. The thickness of the slab achieved with our transformations will no longer be restricted by the evanescent modes and the source can be placed at any position along the boundary of the substrate without affecting the radiation pattern. We also discuss the case where reduced parameters are used,whic h still performs well and is physically realizable.

2008-02-10 PIER Vol. 81, 413-436, 2008. doi:10.2528/PIER07121703

SPM Numerical Results from an Effective Surface Impedance for a One-Dimensional Perfectly-Conducting Rough Sea Surface

Yohann Brelet and Christophe Bourlier

From the analytical theory of rough surface Green's function based on the extension of the diagram method of Bass, Fuks and Itô, with the smoothing approximation, numerical results are presented for Gaussian and sea spectra and compared with a benchmark method by considering a one-dimensional perfectly conducting Gaussian rough surface. The effects of multiple scattering due to the surface roughness are incorporated systematically into the solutions through an effective surface impedance, which can be iterated up to the second-order. In addition, comparisons of the bistatic scattering coefficients are presented with the first- and second- orders conventional small perturbation method. This study will be useful for remote sensing of the ocean surface, especially when the transmitter is close to the surface.

2008-02-06 PIER Vol. 81, 393-412, 2008. doi:10.2528/PIER08012502

Numerical Approach to Survey the Problem of Electromagnetic Scattering from Resistive Strips Based on Using a Set of Orthogonal Basis Functions

Saeed Hatamzadeh, Mohammad Naser-Moghadasi, Esmail Babolian, and Zahra Masouri

In this paper, the problem of electromagnetic scattering from resistive strips is solved and discussed. This problem is modeled by the integral equations of the second kind. The basic mathematical concept is collocation method using block-pulse orthogonal basis functions. An effective numerical method for solving these integral equations is proposed. The problem of electromagnetic scattering from resistive strips is treated in detail, the illustrative computations are given for several cases, and an extensive discussion on the obtained results is performed. This method can be generalized to apply to objects of arbitrary geometry.

2008-02-05 PIER Vol. 81, 371-391, 2008. doi:10.2528/PIER08012303

Electromagnetic Field of a Horizontal Electric Dipole in the Presence of a Four-Layered Region

Yi Hui Xu, Kai Li, and Liang Liu

In this paper, we study in detail the electromagnetic field excited by a horizontal electric dipole in the presence of a four-layered region, which consists of a perfect conductor, the two dielectric layers, and air above. From the derivations and analysis, it is seen that the electromagnetic field includes four wave modes: Direct wave, ideal reflected wave, trapped surface wave, and lateral wave. The wave numbers of the trapped surface wave, which are determined by the residues of the poles, are between the wave number k0 in the air and k2 in the lower dielectric layer. The lateral waves with the wave number being k0 are determined by the integrations along the branch cuts. It should be pointed out that both the trapped surface wave and lateral wave can be separated into the electric-type terms and magnetic-type terms. Analysis and computations show that the trapped surface waves play major roles at large propagation distance when both the dipole point and the observation point are on or close to the air-dielectric boundary.

2008-02-05 PIER Vol. 81, 359-370, 2008. doi:10.2528/PIER08011804

Chip Impedance Matching for UHF RFID Tag Antenna Design

Chye-Hwa Loo, Khaled Elmahgoub, Fan Yang, Atef Elsherbeni, Darko Kajfez, Ahmed Kishk, Tamer Elsherbeni, Leena Ukkonen, Lauri Sydanheimo, Markku Kivikoski, Sari Merilampi, and Pekka Ruuskanen

Passive UHF RFID tag consists of a microchip attached directly to an antenna. Proper impedance match between the antenna and the chip is crucial in RFID tag design. It directly influences RFID system performance characteristics such as the range of a tag. It is known that an RFID microchip is a nonlinear load whose complex impedance in each state varies with the frequency and the input power. This paper illustrates a proper calculation of the tag power reflection coefficient for maximum power transfer by taking into account of the changing chip impedance versus frequency.

2008-02-03 PIER Vol. 81, 343-357, 2008. doi:10.2528/PIER08012402

Investigation of Electromagnetic Complex Scattering for Conductor Target Based on Electromagnetic Images Method

Ying-Le Li, Ji-Ying Huang, and Ming-Jun Wang

The general image relations of electromagnetic sources are presented around a conductor sphere. The general transformations of trigonometric functions and the unit vectors between two coordinates depart from a distance are obtained. The second scattering field for a target is derived in detail. The complex scattering field and the complex RCS are gained respectively. Results show that the electromagnetic interaction of the targets must be calculated as the distance between two targets is small. The second scattering field is small to three order in magnitude to its first scattering field as the distance becomes large. The phase shift of the second field is mainly determined by the target size and the observing position and not affected greatly by its surrounding target and the distance apart. The distortion of a pulse wave is mainly induced by the phase shift of the second scattering field from the particles as the wave propagating through the random discrete medium.

2008-02-03 PIER Vol. 81, 329-341, 2008. doi:10.2528/PIER08011305

Kd-Tree Based Fast Ray Tracing for RCS Prediction

Yu Bo Tao, Hai Lin, and Hu Bao

Ray tracing is of great use for computational electromagnetics, such as the well-known shooting and bouncing ray (SBR) method. In this paper, the kd-tree data structure, coupled with the mailbox technique, is proposed to accelerate the ray tracing in the SBR. The kd-tree is highly effective in handling the irregularly distribution of patches of the target, while the repeatedly intersection tests between the ray and the patch when using space division acceleration structures can be eliminated through the mailbox technique. Numerical results show excellent agreement with the measured data and the exact solution, and demonstrate that the kd-tree as well as the mailbox technique can greatly reduce the computation time.

2008-01-31 PIER Vol. 81, 315-328, 2008. doi:10.2528/PIER08010202

Robust Adaptive Beamforming for a Class of Gaussian Steering Vector Mismatch

Yujie Gu, Zhi-Guo Shi, Kang Chen, and Yu Li

Compared with the worst-case optimization-based approach, the probability-constrained approach is a more flexible one to robust adaptive beamforming. In this paper, a precise relationship between the two approaches is built in the case of Gaussian steering vector mismatch, which shows that the probability-constrained beamformer design can be interpreted in terms of the worst-case beamformer design. Numerical simulations demonstrate that the precise version of the probability-constrained beamformer is more robust to the steering vector mismatch than the other popular robust adaptive beamformers.

2008-01-30 PIER Vol. 81, 303-314, 2008. doi:10.2528/PIER08011603

Adaptively Accelerated GMRES Fast Fourier Transform Method for Electromagnetic Scattering

Yuan Xin and Ping-Liang Rui

The problem of electromagnetic scattering by 3D dielectric bodies is formulated in terms of a weak-form volume integral equation. Applying Galerkin's method with rooftop functions as basis and testing functions,the integral equation can be usually solved by Krylov-subspace fast Fourier transform (FFT) iterative methods. In this paper,the generalized minimum residual (GMRES)-FFT method is used to solve this integral equation,and several adaptive acceleration techniques are proposed to improve the convergence rate of the GMRES-FFT method. On several electromagnetic scattering problems,the performance of these adaptively accelerated GMRESFFT methods are thoroughly analyzed and compared.

2008-01-26 PIER Vol. 81, 291-302, 2008. doi:10.2528/PIER08012003

Multilevel Fast Multipole Algorithm for Radiation Characteristics of Shipborne Antennas Above Seawater

Xun-Wang Zhao, Chang-Hong Liang, and Le Liang

Radiation characteristics of shipborne antennas above lossy half-space are studied using the multilevel fast multipole algorithm (MLFMA). The near terms in the MLFMA are evaluated by using the rigorous half-space dyadic Green's function, computed via the method of complex images. The far MLFMA interactions employ an approximate dyadic Green's function via a direct-radiation term plus a single real image, with the image amplitude characterized by the polarization-dependent Fresnel reflection coefficient. Finally, radiation patterns of an ultra-shortwave antenna mounted on a realistic 3-D ship over seawater are presented and compared with a rigorous method-ofmoments (MoM) solution.

2008-01-23 PIER Vol. 81, 279-289, 2008. doi:10.2528/PIER08011402

Indoor Accurate RCS Measurement Technique on UHF Band

Chu-Feng Hu, Jia-Dong Xu, N. J. Li, and Lin-Xi Zhang

Based on the step-frequency RCS measurement system, high performance absorbers and low scattering supports, employing two log-periodic dipole antennas to carry out the quasi-monostatic measurement and many DSP techniques to reduce the error, the indoor accurate RCS measurement can be completed on UHF band. Experimental results show that the valid data waved less than 1 dB can be obtained over 70% of whole band.

2008-01-23 PIER Vol. 81, 267-277, 2008. doi:10.2528/PIER08011104

Transition Behavior of k -Surface: from Hyperbola to Ellipse

Shan Qiao, Guoan Zheng, Haifei Zhang, and Li-Xin Ran

The transition behavior of the k-surface of a lossy anisotropic indefinite slab is investigated. It is found that, if the material loss is taken into account, the k-surface does not show a sudden change from hyperbola to the ellipse when one principle element of the permittivity tensor changes from negative to positive. In fact, after introducing a small material loss, the shape of the k-surface can be a combination of a hyperbola and an ellipse, and a selective high directional transmission can be obtained in such a slab.

2008-01-22 PIER Vol. 81, 253-265, 2008. doi:10.2528/PIER07122602

FDTD Study on Wave Propagation in Layered Structures with Biaxial Anisotropic Metamaterials

Mao-Yan Wang, Jun Xu, Jian Wu, Bing Wei, Hai-Long Li, Tong Xu, and De-Biao Ge

The Gaussian beam propagation in multi-layered structures that include indefinite anisotropic metamaterial (AMM) are simulated with shift operator method in Finite-Difference Time-Domain method (FDTD). The excitations of backward and forward surface affected by the types of biaxial AMM are investigated. Numerical results show that the directions of the guided wave excited are influenced by the sign of z component of relative permeability tensor of AMM that determines the energy flow is positively refracted or negatively refracted. Positive or negative Goos-H¨anchen shift associated with Total Cutoff media are also shown.

2008-01-22 PIER Vol. 81, 237-252, 2008. doi:10.2528/PIER08011102

Analysis of Wave Function, Energy and Transmission Coefficients in GaN /AlGaN Superlattice Nanostructures

Kanchan Talele and Dyneshwar Patil

Analysis of wave function intensity, eigen energy and transmission coefficients in GaN/AlGaN superlattice nanostructure has been carried out using Transfer Matrix Method (TMM). The effect of change in Aluminum mole fraction in AlxGa1xN barrier region has been included through variable effective mass in the Schrödinger time independent equation. The behaviour of wave function intensity has been studied for superlattice structure by changing the barrier width. The effect of smaller barrier width on wave function intensity in case of superlattice is clearly observed due to interaction of wave functions in the adjacent wells and it provides a new insight in the nature of interacting wave functions for thin barriers in GaN/AlGaN superlattice structures. The barrier widths have been optimized for the varying number of wells leading to better quantum confinement. The iterative method (Secant Method) is used to determine value of electron energy E. The number of iterations need to converge the value of E has been simulated. Transmission coefficients have been determined as a function of energy E considering tunneling effect for three well structures using TMM. Analysis has been extended to show surface image of wave function intensity for 5 and 6 wells.

2008-01-19 PIER Vol. 81, 225-235, 2008. doi:10.2528/PIER08010902

Design of Photonic Band Gap Filter

Ragini Srivastava, Khem Thapa, Shyam Pati, and Sant Ojha

In this paper a new type of optical filter using photonic band gap materials has been suggested. A detailed mathematical analysis is presented to predict allowed and forbidden bands of wavelengths with variation of angle of incidence. It is possible to get desired ranges of the electromagnetic spectrum filtered with this structure by changing the incidence angle of light.

2008-01-19 PIER Vol. 81, 213-223, 2008. doi:10.2528/PIER08010403

Improved Channel Estimation and MAI-Robust Schemes for Wireless Ofdma System

Y.-D. Lee, D.-H. Park, and Hyoung-Kyu Song

This paper presents an investigation into improving the channel estimation scheme and reducing the effects of symbol timing misalignment when OFDMA is used as an access scheme. Under OFDMA uplink channel environments, appropriate symbol length of CAZAC sequences as a preamble could be utilized in accordance with the number of transmitting antenna and channel condition. The effect of the number of CAZAC sequences for channel estimation is also presented in terms of mean square error (MSE). Taking into account the effect of multiple access interference (MAI) introduced by a symbol timing misalignment, the symbol error rate (BER) and throughput performance are investigated for a typical OFDMA uplink scenario.

2008-01-19 PIER Vol. 81, 197-212, 2008. doi:10.2528/PIER08010303

Numerical Simulation of Propagation of EM Pulse through Lossless Non-Uniform Dielectric Slab Using Characteristic-Based Method

Mingtsu Ho, Fu-Shun Lai, Shun-Wen Tan, and Pi-Wei Chen

This paper demonstrates the one-dimensional computational results of the propagation of Gaussian electromagnetic pulse through dielectric slabs of finite thickness with variation in permittivity. The numerical approach used is the characteristic-based method solving the time-domain Maxwell curl equations involved with nonuniform permittivity. In the numerical model, all dielectric slabs are assumed to be isotropic, lossless, and linear. The permittivity of dielectric slab may increase or decrease linearly or sinusoidally. The numerical permittivity is finely discretized such that the variation between two adjacent grids is so small that the non-uniform permittivity is assumed to be piecewise continuous and consequently can be modeled as an individual block. The numerical results of various electric fields, both in the time- and frequency-domain, are presented and compared based on the dielectric slab of constant permittivity for close investigating the effects of the non-uniform permittivity distribution on the electromagnetic fields. It is also shown that under certain arrangement of Gaussian electromagnetic pulse and dielectric slab thickness the pattern of field propagation, reflection and transmission, can be reproduced in different time scales and frequency ranges.

2008-01-19 PIER Vol. 81, 183-195, 2008. doi:10.2528/PIER08010103

A New 2-18 GHz Quad-Ridged Horn Antenna

Ramin Dehdasht-Heydari, Hamid Reza Hassani, and Ali Reza Mallahzadeh

A novel design of a dual-polarized broadband 2-18 GHz horn antenna with VSWR≤2.2 is presented. The designed horn antenna is most suitable as a feed element in reflectors of the radar systems and EMC applications. A coaxial line to quadruple-ridged waveguide transition with a new conical cavity back and a technique for tapering the flared section of the horn is introduced to improve the return loss and matching of the impedance, respectively. In order to overcome the deterioration of the broadside radiation pattern at higher frequencies, common to broad band ridged horn antennas, a new modified horn antenna with arc shaped aperture is introduced. Results of simulation obtained via two different software packages, HFSS and CST, for VSWR, isolation, radiation patterns, and gain of the designed quad ridged horn antenna as well as the modified horn antenna are presented and discussed.

2008-01-17 PIER Vol. 81, 167-182, 2008. doi:10.2528/PIER07122601

A Hybrid Multimode Contour Integral Method for Analysis of the h -Plane Waveguide Discontinuties

Ali Banai and Ali Hashemi

A hybrid method is introduced for analysis of the H-plane waveguide discontinuities. It combines multimode contour integral and mode matching techniques. The process is based on dividing the circuit structure into key building blocks and finding the multimode scattering matrix of each block individually. The multimode scattering matrix of the whole structure can be found by cascading these blocks. Also contour integral method is developed for analysis of multi-media circuits. Therefore, it is possible to analyze H-plane waveguide filters with dielectric resonators using this method. The accuracy and run time of the purposed method is compared with those reported in literatures and/or Ansoft HFSS software.

2008-01-17 PIER Vol. 81, 149-166, 2008. doi:10.2528/PIER07121801

Stencil Coefficient Computations for the Multiresolution Time Domain Method - a Filterbank Approach

Sevoor Meenakshisund Vaitheeswaran and S. Narasimhan

Multiresolution Time Domain (MRTD) techniques based on wavelet expansions can be used for adaptive refinement of computations to economize the resources in regions of space and time where the fields or circuit parameters or their derivatives are large. Hitherto, standard wavelets filter coefficients have been used with the MRTD method but the design of such filter itself may enable to incorporate desired properties for different applications. Towards this, in this paper, a new set of stencil coefficients in terms of scaling coefficients starting from a half band filter, designed by window method and deriving a physically realizable filter by spectral factorization using cepstral technique, for the MRTD method is presented. These stencil coefficients for the MRTD are found to give good agreement with similar MRTD schemes such as those obtained using Daubechies orthogonal wavelets.

2008-01-17 PIER Vol. 81, 135-148, 2008. doi:10.2528/PIER08010901

High-Frequency Expression for the Field in the Caustic Region of a PEMC Gregorian System Using Maslov's Method

Muhammad Fiaz, Abdul Aziz, Abdul Ghaffar, and Qaisar Naqvi

Method proposed by Maslov has been used, to remedy the problem of geometrical optics, for a two dimensional Perfect electromagnetic conductor (PEMC) Gregorian system. It generates an integral form of solution near the caustic that can be evaluated analytically/numerically, or withuniform asymptotic techniques. Away from the caustic it recovers the geometrical optics field. Numerical computations are made to calculate the field around the caustic of a Gregorian system.

2008-01-17 PIER Vol. 81, 121-133, 2008. doi:10.2528/PIER08010206

Radar Target Recognition by Projected Features of Frequency-Diversity RCS

Kun-Chou Lee, Chih-Wei Huang, and Ming-Chung Fang

In this paper, the radar target recognition is given by projected features of frequency-diversity RCS (radar cross section). The frequency diversity means signals are collected by sweeping the frequency of the incident illumination. Initially, the frequencydiversity RCS data from targets are collected and projected onto the PCA (principal components analysis) space. The elementary recognition of targets is efficiently performed on the PCA space. To achieve well separate recognition of targets, the features of the PCA space are further projected onto the LDA (linear discriminant algorithm) space. Simulation results show that accurate results of radar target recognition can be obtained by the proposed frequencydiversity scheme. In addition, the proposed frequency-diversity scheme has good ability to tolerate noise effects in radar target recognition.

2008-01-08 PIER Vol. 81, 99-119, 2008. doi:10.2528/PIER07122502

A Moment Method Simulation of Electromagnetic Scattering from Conducting Bodies

Saeed Hatamzadeh, Mohammad Naser-Moghadasi, and Zahra Masouri

In this paper a moment method simulation of electromagnetic scattering problem is presented. An effective numerical method for solving this problem based on the method of moments and using block-pulse basis functions is proposed. Some examples of engineering interest are included to illustrate the procedure. The scattering problem is treated in detail, and illustrative computations are given for some cases. This method can be generalized to apply to objects of arbitrary geometry and arbitrary material.

2008-01-08 PIER Vol. 81, 73-97, 2008. doi:10.2528/PIER07122404

Truncated Cosine Fourier Series Expansion Method for Solving 2-D Inverse Scattering Problems

Abbas Semnani and Manouchehr Kamyab

Truncated cosine Fourier series expansion method is applied for reconstruction of lossy and inhomogeneous 2-D media by using inverse scattering method in time domain. In this method, the unknown parameters are expanded in a cosine Fourier series and coefficients of this expansion are optimized in particle swarm optimization (PSO) routine with the aid of finite difference time domain (FDTD) method as an electromagnetic (EM) solver. The performance of the algorithm is studied for several 2-D permittivity and conductivity profile reconstruction cases. It is shown that since only a limited number of terms are retained in the expansion, using the proposed method guarantees the well-posedness of the problem and uniqueness of the solution and various types of regularization may be used to only have more precise reconstruction. It is also shown that the number of unknowns in optimization routine is reduced more than 75 percent as compared with conventional methods which leads to a considerable reduction in the amount of computations with negligible adverse effect on the precision of reconstruction. Sensitivity analysis of the suggested method to the number of expansion terms in the algorithm is studied, as well.

2008-01-08 PIER Vol. 81, 61-72, 2008. doi:10.2528/PIER07121704

Characterization of the Susceptibility of Integrated Circuits with Induction Caused by High Power Microwaves

Sun-Mook Hwang, Joo-Il Hong, and Chang-Su Huh

This paper examines malfunction and destruction of semiconductors by high power microwaves. The experiments employ a waveguide and a magnetron to study the influence of high power microwaves on TTL/CMOS IC inverters. The TTL/CMOS IC inverters are composed of a LED circuit for visual discernment. A CMOS IC inverter damaged by a high power microwave is observed with power supply current and delay time. When the power supply current was increased 2.14times for normal current at 10 kV/m, the CMOS inverter was broken by latch-up. The CMOS inverter damaged by latch-up returned its original level of functioning, because parasitic impedance inside the chip increased with the elapse of time. Three different types of damage were observed by microscopic analysis: component, onchipwire, and bondwire destruction. Based on the results, TTL/CMOS IC inverters can be applied to database to elucidate the effects of microwaves on electronic equipment.

2008-01-14 PIER Vol. 81, 41-59, 2008. doi:10.2528/PIER07120402

A Novel Method for Testing Ultra Wideband Antenna-Feeds on Radio Telescope Dish Antennas

Shubhendu Joardar and Ashit Bhattacharya

A prototype of dual polarized ultra wideband antenna-feed was fabricated (200 MHz-2000 MHz) for usage in low frequency radio astronomy at the GMRT (Giant Meterwave Radio Telescope). The feed was mounted on one of the GMRT dish antennas and tested using signals from the radio galaxy Cygnus A and zenith sky. Modeling of the GMRT dish with the antenna-feed mounted at its focus has been done. With the use of simulated radiation patterns of the antennafeed and the measured results, the system independent deflections at various frequencies were computed. The performances of the GMRT dish with present mesh and RMS (root mean square) efficiencies have been studied. Performance of the GMRT dish with 100% mesh and RMS efficiencies were computed. Future guidelines of ultra wideband antenna-feed design for GMRT have been indicated. Improvements of the GMRT dishes have also been suggested.

2008-01-03 PIER Vol. 81, 21-39, 2008. doi:10.2528/PIER07121502

Vector Finite Element Analysis of Multicomponent Induction Response in Anisotropic Formations

Xiang Sun and Zai-Ping Nie

Multicomponent induction logging responses are simulated by using hierarchical mixed order vector finite element method (FEM). In order to modeling three orthogonal magnetic dipoles, we adopt the method that the total field is separated into incident field and secondary field, and only the secondary field is computed by FEM. In addition, two techniques are applied to improve the modeling accuracy and computational efficiency: 1) Hierarchical mixed order vector basis functions are applied to FEM. Different order basis functions are used in different elements in accordance with the changing speed of the field. The mixed order scheme reduces greatly the number of unknowns without reducing accuracy, and can attain much higher computational efficiency. 2) The systemof the FEM equations is solved by Distributed-SuperLU, and the results of multiple measure points can be got simultaneously. The FEM result is validated against volume integral equation method and the approach of planar layered media Green's functions, and the comparisons show very good agreement. Finally, the multicomponent induction response in anisotropic formations involving eccentric tools and dipping beds is included to demonstrate the flexibility of the method.

2008-01-03 PIER Vol. 81, 1-19, 2008. doi:10.2528/PIER07121302

From Cpu to GPU: GPU-Based Electromagnetic Computing (GPUECO)

Yu Bo Tao, Hai Lin, and Hu Bao

In this paper, we provide a new architecture by using the programmable graphics processing unit (GPU) to move all electromagnetic computing code to graphical hardware, which significantly accelerates Graphical electromagnetic computing (GRECO) method. We name this method GPUECO. The GPUECO method not only employs the hidden surface removal technique of graphics hardware to identify the surfaces and wedges visible from the radar direction, but also utilizes the formidable of computing power in programmable GPUs to predict the scattered fields of visible surfaces and wedges using the Physical Optical (PO) and Equivalent Edge Current (EEC). The computational efficiency of the scattered field in fragment processors is further enhanced using the Z-Cull and parallel reduction techniques, which avoid the inconsistent branching and the addition of the scattered fields in CPU, respectively. Numerical results show excellent agreement with the exact solution and measured data and, the GPUECO method yields approximately 30times faster results.