1,721,111 research outputs found
Singular value optimization in inverse electromagnetic scattering
Full text linkWe provide an approach to reformulate the problem of extracting the information on scatterers in inverse problems. The approach is based on the recently introduced concept of singular value optimization and exploits the single scattering hypothesis of the Born approximation. A two-dimensional geometry with multifrequency/single-view plane wave illumination and TM polarization with near-field acquisitions is considered. Numerical results show how it is possible to dramatically reduce the number of near-field samples against the canonical λm in/2 criterion. © 2016 IEEE
FAST GPU-BASED INTERPOLATION FOR SAR BACK- PROJECTION
No full textWe introduce and discuss a parallel SAR backprojection algorithm using a Non-Uniform FFT (NUFFT) routine implemented on a GPU in CUDA language. The details of a convenient GPU implementation of the NUFFT-based SAR backprojection algorithm, amenable to further generalizations to a multi-GPU architecture, are also given. The performance of the approach is analyzed in terms of accuracy and computational speed by comparisons to a ``standard", parallel version of the backprojection algorithm exploiting FFT+interpolation instead of the NUFFT. Different interpolators have been considered for the latter processing scheme. The NUFFT-based backprojection has proven significantly more accurate than all the compared approach, with a computing time of the same order. An analysis of the computational burden of all the different steps involved in both the considered approaches (i.e., standard and NUFFT backprojections) has been also reported. Experimental results against the Air Force Research Laboratory (AFRL) airborne data delivered under the ``challenge problem for SAR-based Ground Moving Target Identification (GMTI) in urban environments" and collected over circular flight paths are also shown
Multi-resolution imaging with an optimized number and distribution of sampling points
No full textWe propose an approach of interest in Imaging and Synthetic Aperture Radar (SAR) tomography, for the optimal determination of the scanning region dimension, of the number of sampling points therein, and their spatial distribution, in the case of single frequency monostatic multi-view and multi-static single-view target reflectivity reconstruction. The method recasts the reconstruction of the target reflectivity from the field data collected on the scanning region in terms of a finite dimensional algebraic linear inverse problem. The dimension of the scanning region, the number and the positions of the sampling points are optimally determined by optimizing the singular value behavior of the matrix defining the linear operator. Single resolution, multi-resolution and dynamic multi-resolution can be afforded by the method, allowing a flexibility not available in previous approaches. The performance has been evaluated via a numerical and experimental analysis
Gradient-based near-field antenna characterization in planar geometry
Full text linkA near field characterization technique, based on the optimization of the Singular Value Behavior, is here accelerated by using the analytical expression of the gradient of the relevant objective functional. This solution allows to conveniently tackle the cases of electrically large antennas
An efficient approach to the near-field sampling of electromagnetic fields
No full textThe paper presents a novel and efficient approach to near-field planar quasi-raster acquisitions. The method outperforms existing techniques in terms of number of field samples and acquisition time. The performance is shown with an experimental test
Field sampling of incoherent sources
No full textThe problem of characterizing random sources from near-field measurements performed on a domain DI and of devising the random field sampling procedure is tackled by a stochastic approach. The presented technique is an extension of that introduced in [A. Capozzoli, et al., Field sampling and field reconstruction: a new perspective, Radio Sci., vol. 45, 2010] and successfully adopted to experimentally characterize deterministic (CW) radiators and fields. Under the assumption that the source is wide sense stationary, quasi-monochromatic and incoherent, its intensity is reconstructed by time-domain field measurements aimed at extracting information from the mutual coherence of the field on DI. The linear relation between the field coherence on DI and the source intensity is inverted by using the Singular Value Decomposition (SVD) approach, properly representing the source intensity distribution by exploiting the a priori information (e.g., its size and shape) on the radiator. The sampling of the radiated random field is devised by a singular value optimization procedure of the relevant linear operator. Numerical results sketch the performance of the approach
Multi-frequency planar near-field scanning by means of singular-value decomposition (SVD) optimization
No full textAn optimized near-field/far-field (NFFF) transformation for characterizing planar aperture antennas from “quasi-raster” and plane-polar multi-frequency scanning data is presented. The method is a generalization of that introduced by the Authors and then used for plane-polar and “quasi-raster” scans, respectively. The generalization consists of characterizing antennas at different frequencies by defining a common frequency-measurement grid to save scanning time. The method tackles the multi-frequency measurement problem by a linear operator, A, and solves the problem as a singular-value optimization of A. The field-sample positions are then chosen to provide the minimum number of near-field samples optimizing the singular-value dynamics of A. The computational burden is dealt with by proper programming on graphics processing units (GPUs). Numerical and experimental results show the effectiveness of the technique
Resolution Control for SAR Tomography with Optimized Track Distribution
No full textSAR tomography [1,2] allows accessing 3D information on the investigated scenario by constructing a synthetic aperture transverse to the flight direction by exploiting several passes of the same sensor over the same scene or using sensors constellations working simultaneously.
Indeed, the amount of information that can be acquired in the height direction is limited by the difficulties and costs of realizing many passes of the same sensor on the same scene, including the effects of temporal decorrelation, or alternatively of implementing constellations of several sensors.
Operating with “few” acquisitions requires defining the measurement constellation maximizing the collected information, given some knowledge on the scene of interest. This problem, which arises to be crucial in defining the actual potentialities of SAR tomography, has been generally overlooked until very recently [3-6]. In particular, in [3,4], it has been shown how, by properly exploiting the available a priori information through an appropriate representation of the problem unknowns and by a suitable optimization of the measurement constellation, the number of required acquisitions can be reduced.
The purpose of this paper is to extend the technique in [3,4] so that the number and the (generally irregular) positions of the tracks to be flown can be flexibly linked both to the a priori information on the investigated scenario, and to the information one is actually interested to extract about the scene at hand. This includes the possibility to handling the number and positions of the tracks as a function of desired degree of resolution/multiresolution [7].
To this end, the reflectivity distribution of the scene is represented by prolate spheroidal wave functions, properly filtered to a desired resolution limit [7], and the measurement constellation is designed in terms of track number and positions, by a singular value optimization procedure [8]. A numerical analysis shows the potentialities of the technique
Fast and Accurate Processing for Magnetic Resonance Imaging under non-Cartesian acquisitions using NUFFTs and GPUs
No full textWe present the development of a fully parallel Magnetic Resonance Imaging processing written in CUDA language capable to very quickly process data collected under non-conventional (e.g., spiral) acquisition schemes
THE DESIGN OF AN OPTICAL TIME STEERED ANTENNA BASED ON A NEW INTEGRATED TRUE TIME DELAY UNIT
No full textIn the framework of wide-band and ultra wide-band array antennas, an Optical Time Steered Antenna (OTSA) is presented, by considering the design strategies of a new True Time Delay (TTD) Control Unit in the Beam Forming Network (BFN). The Unit has high reliability, low crosstalk, low switching time and a potential low cost, being based on a low cost technology. Furthermore, due to its compactness and modularity, can be easily grouped with other ones to make a control unit of large arrays. Different strategies and working configurations of the TTD control unit are presented as a trade- off among hardware complexity, insertion loss reduction and beam control capability. The design of an OTSA prototype is discussed by considering a realistic model simulating the behavior of a real world antenna and accounting for unavoidable non-realities, such as random, periodic and systematic errors introduced by each device exploited in the OTSA as well as mutual coupling between radiating elements. An optimal trimming strategy, able to compensate at best for BFN errors and based on the use of suitably located trimmers, is presented. Among other cases, to enlighten the potentialities of the OTSA, an all optical architecture providing a difference beam squint free pattern is also proposed
- …
