Borja Gonzalez-Valdes

3D Whole Body Imaging for Detecting Explosive-Related Threats

In this work, a whole-body imaging system for concealed objects detection using a mm-wave radar is presented. 3D high resolution images are generated using a two step process. Initially, an inverse source-based Fast Multipole Method (iFMM) provides a first approximation to the true human torso. Afterwards, the retrieved geometry is refined using the Iterative Field Matrix (IFM) technique. Assuming smooth variations of the human body profile, the object detection is performed by comparing the retrieved surface with a smoothed one. Results are based on Physical Optics simulations of the human body, considering both cases with and without objects.

Zooming and Scanning Gregorian Confocal Dual Reflector Antennas

The zooming and scanning capabilities of a Gregorian confocal dual reflector antenna are described. The basic antenna configuration consists of two oppositely facing paraboloidal reflectors sharing a common focal point. A planar feed array is used to illuminate the subreflector allowing the antenna to scan its beam. The resulting quadratic aberrations can be compensated by active mechanical deformation of the subreflector surface, which is based on translation, rotation and focal length adjustment. In order to reduce the complexity of the mechanical deformation, least squares fit paraboloids are defined to approximate the optimal correction surface. These best fit paraboloids considerably reduce scanning losses and pattern degradation. This work also introduces two different zooming techniques for the Gregorian confocal dual reflector antenna: the first consists of introducing a controlled quadratic path error to the main reflector aperture; and the second is based on reducing the size of the radiating aperture of the feeding array.

On the Use of Compressed Sensing Techniques for Improving Multistatic Millimeter-Wave Portal-Based Personnel Screening

This work develops compressed sensing techniques to improve the performance of an active three dimensional (3D) millimeter wave imaging system for personnel security screening. The system is able to produce a high-resolution 3D reconstruction of the whole human body surface and reveal concealed objects under clothing. Innovative multistatic millimeter wave radar designs and algorithms, which have been previously validated, are combined to improve the reconstruction results over previous approaches. Compressed Sensing techniques are used to drastically reduce the number of sensors, thus simplifying the system design and fabrication. Representative simulation results showing good performance of the proposed system are provided and supported by several sample measurements.

Sparse Array Optimization Using Simulated Annealing and Compressed Sensing for Near-Field Millimeter Wave Imaging

The optimization and use of a sparse array configuration for an active three dimensional (3D) millimeter wave imaging system for personnel security screening is presented in this work. The combination of the optimization procedure with the use of Compressed Sensing techniques allows drastic reduction in the number of sensors, thereby simplifying the system design and fabrication and reducing its cost. Representative simulation results showing good performance of the proposed system are provided and supported by sample measurements.

Fourier-Based Imaging for Multistatic Radar Systems

Fourier-based methods for monostatic and bistatic setups have been widely used for high-accuracy radar imaging. However, the multistatic configuration has several characteristics that make Fourier processing more challenging: 1) a nonuniform grid in k-space, which requires multidimensional interpolation methods, and 2) image distortion when the incident spherical wave is approximated by a plane wave. This contribution presents a Fourier-based imaging method for multistatic systems, solving the aforementioned limitations: the first, by using k-space partitioning and applying interpolation in each domain; the second, by approximating the spherical wave with multiple plane waves. Both solutions are fully parallelizable, thus allowing calculation time savings. Validation and benchmarking with a synthetic aperture radar backpropagation algorithm have been performed through 2-D and 3-D simulation-based examples. Imaging results from radar measurements have been assessed.

A Bifocal Ellipsoidal Gregorian Reflector System for THz Imaging Applications

Current terahertz imagers rely on reflector systems for the beam quality and imaging speed because the cross-range span that the system can cover is limited by the beam aberrations when the antenna scans. We present the design of a Bifocal reflector system that can rapidly scan a terahertz beam for standoff imaging applications while increasing the field of view of previous designs up to 50%. The design is based in a confocal Gregorian system where the nominal reflector surfaces are substituted by shaped surfaces to reduce the beam aberrations, while not increasing the manufacture cost of the reflector antenna. We also provide a set of useful design formulas for the design of this kind of reflector systems. The beam patterns obtained by the proposed designs are numerically calculated with the commercial software GRASP and compared with those obtained with previous approaches to the same problem, showing the better performance of the proposed solution.

Reconstructing Distortions on Reflector Antennas With the Iterative-Field-Matrix Method Using Near-Field Observation Data

This work extends the mathematical formulation of the iterative-field-matrix method for observed data from the near-field region of a Perfect Electric Conductor. The method is used as a diagnosis tool for reflector antennas, to determine the positions and extent of distortions from their idealized shapes. The new formulation is tested on a reflector antenna with several significant bumps, and excellent results are achieved. This work also presents an example where the Method of Moments is used to generate the synthetic data and the inversion is performed using Physical Optics. Such a configuration ensures that the forward model is unbiased with respect to the inversion model, demonstrating that the new formulation is also robust for these realistic scenarios.

3-D High-Resolution Imaging Radar at 300 GHz With Enhanced FoV

We have developed a 3-D high-resolution radar at 300 GHz with a cell resolution of 1×1.6×1.6 cm3 at a standoff distance of 8 m for security applications. The selection of an operating frequency of 300 GHz, a bandwidth of 27 GHz, and a field of view of 50×90 cm2 improves the through-clothes imaging of person-borne concealed objects. The radars antenna design allows single-pixel imaging at a frame rate of two frames per second. A reduction in cost and power consumption and improvements in image quality are additional requirements taken into account in the design of this radar system.

On the Use of Improved Imaging Techniques for the Development of a Multistatic Three-Dimensional Millimeter-Wave Portal for Personnel Screening

The design and evaluation of an active three dimensional (3D) millimeter wave imaging system for personnel security screening is presented in this work. The system is able to produce a high- resolution 3D reconstruction of the whole human body surface and reveal concealed objects under clothing. Innovative multistatic millimeter wave radar designs and algorithms, which have been previously validated, are combined to signiflcantly improve the previous reconstruction results. In addition, the system makes use of a reduced amount of information, thus simplifying portal design. Representative simulation results showing good performance of the proposed system are provided and supported by sample measurements.

Generating Contoured Beams With Single-Shaped Reflectors Using a Iterative Field-Matrix Approach

In this letter, a new method is presented to synthesize contoured beams by specifically shaping a single reflector antenna. The algorithm is based on the iterative-field-matrix (IFM) method, which finds a linear relationship, in a matrix form, between a local surface distortion and the difference of the nominal and objective far-field patterns. The matrix is inverted using singular value decomposition (SVD) and Tikhonov regularization; as a result, fast convergence is achieved. The method has been successfully applied to synthesize contoured beams for two different scenarios.