UWB target classification using SVM

Bouza, Magdalena; Altieri, Andrés; Galarza, Cecilia G.

doi:10.1109/argencon.2018.8646072

Cited by 3 publications

(3 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For this purpose, we observe objects of indistinguishable shapes composed of different materials. This makes for a more difficult problem since variations arising from shape differences are easier to detect than those produced by differences in composition [27]. In both cases, we compared the results with those obtained by processing time-domain signals.…”

Section: Introduction and Main Contributionsmentioning

confidence: 99%

Robust Target Classification Using UWB Sensing^*

2023

Self Cite

View full text Add to dashboard Cite

Contactless material characterization has received widespread attention in the radar and engineering domains. Specifically, impulsive Ultra Wideband (UWB) systems are a versatile technology for the nondestructive characterization of samples because the scattered field produced by the targets is highly dependent on their composition and shape. After the initial transient response to the transmitted pulse, the scattered signal can be decomposed as a sum of complex exponentials, called complex natural resonances (CNR), which are dependent only on the geometry and composition of the target. Using this result, a classification problem was formulated to discriminate among targets, and a processing strategy was proposed to solve it. In particular, by using spectral decomposition tools, the information obtained from the physical model can be exploited in combination with data-driven learning techniques. Consequently, a classification strategy that is robust to modeling uncertainties and experimental perturbations was designed. To assess the performance of the new scheme, it was tested using both synthetic and experimental data obtained from targets illuminated with a UWB radar. The results showed substantial gains compared to classification using time-domain signals. INDEX TERMSComplex natural resonances, pattern classification, radar signal processing, spectral analysis.

show abstract

Section: Introduction and Main Contributionsmentioning

confidence: 99%

Robust Target Classification Using UWB Sensing^*

2023

Self Cite

View full text Add to dashboard Cite

show abstract

“…to other properties, permittivity (or dielectric constant) is promising for the implementation of a sensing unit and offers a potential for various applications in medical, biological, and agricultural fields [3]- [6]. Ultra-Wideband (UWB) radar signals are characterized for having both high frequency carrier and high bandwidth [7]. This makes the scattered field from the targets when irradiated with UWB pulses highly dependent on the composition and shape of the target, the absorption and scattering properties of the material at the wavelengths used, the refractive index, and thus the specular reflection from the material.…”

Section: Introductionmentioning

confidence: 99%

“…al. [7] demonstrated the feasibility of using a UWB pulse radar with a center frequency of 4.25 GHz to discriminate between groups of 3, 4, and 5 materials using a support vector machine classifier. Unlike previous work, Agresti and Milani [8] employed a portable 3D imaging radar-based system (the Walabot sensor by Vayyar Imaging working in the 6.3-8 GHz frequency range) to acquire three-dimensional radiance maps of various analyzed objects and processed that by using CNNs in order to identify which material the object is made of.…”

Section: Introductionmentioning

confidence: 99%

Radar-based Materials Classification Using Deep Wavelet Scattering Transform: A Comparison of Centimeter vs. Millimeter Wave Units

Khushaba,

Hill

2022

Preprint

View full text Add to dashboard Cite

Radar-based materials detection received significant attention in recent years for its potential inclusion in consumer and industrial applications like object recognition for grasping and manufacturing quality assurance and control. Several radar publications were developed for material classification under controlled settings with specific materials' properties and shapes. Recent literature has challenged the earlier findings on radarsbased materials classification claiming that earlier solutions are not easily scaled to industrial applications due to issues such as the analog-to-digital converters' high sensitivity to target aspect angle, noise fluctuations due to temperature and other external conditions and sensor orientation. Published experiments on the impact of the aforementioned factors on the robustness of the extracted radar-based traditional features have already demonstrated that the application of deep neural networks can mitigate, to some extent, the impact to produce a viable solution. However, previous studies lacked an investigation of the usefulness of lower frequency radar units, specifically < 10GHz, against the higher range units around and above 60GHz. To address the aforementioned investigation, this research considers two radar units with different frequency ranges: the Walabot-3D (6.3-8 GHz) cm-wave and IMAGEVK-74 (62-69 GHz) mm-wave imaging units by Vayyar Imaging. A comparison is presented on the applicability of each unit for material classification. This work also extends upon previous efforts, by applying deep wavelet scattering transform for the identification of different materials based on the reflected signals received by these units. In the wavelet scattering feature extractor, data is propagated through a series of wavelet transforms, nonlinearities, and averaging to produce low-variance representations of the reflected radar signals. This work is unique in terms of the comparison of the utilized radar units and algorithms in material classification and includes real-time demonstrations that show strong performance by both units, with increased robustness offered by the cm-wave radar unit.

show abstract