Search

In this paper, an approach for ground-moving target classification with an FMCW radar is proposed. In particular, data are collected using a low-cost 24 GHz off-the-shelf FMCW radar, combined with an embedded Raspberry Pi device for data acquisition and processing. An FFT-based processing scheme is then applied to obtain a sequence of range-Doppler maps, which are provided in input to different convolutional neural network (CNN) architectures for classifying the targets (cars, motorcycles, or pedestrians) eventually passing in front of the radar. Specifically, two approaches have been followed and compared. In the first one, single range-Doppler maps are processed alone using a convolutional neural network, and then a voting mechanism is applied to select the target classes. In the second approach, a sequence of range-Doppler maps is processed using a time-distributed layer feeding a recurrent neural network. The CNNs are deployed on the Raspberry Pi providing the target classification on a low-cost embedded device. The obtained results show that the proposed approaches allow for effectively detecting the different types of targets running on an embedded device in less than one second.

An improved spectral coefficient method based on Chebyshev polynomials of the second kind is employed to solve the vector radiative transfer (VRT) equation under the assumption of parallel-plane atmosphere and Rayleigh scattering. The solver is extended to consider a Lambertian surface at the bottom of the atmosphere (BOA). The computational properties of the proposed algorithm are analyzed, and the validity of the implemented method is tested against a publicly available benchmark dataset.

Magnetic biomaterials are multifunctional tools currently under investigation as theranostic platforms for biomedical applications. They can be implanted in bone tissue after bone cancer resection to perform local interstitial hyperthermia treatment. Given the requirements of high quality treatment, the hyperthermia therapy should be performed monitoring the system temperature, to avoid hot spots and control the treatment outcome. It is known that the magnetic properties of such implants vary with temperature. It is hypotesized that the treatment dynamic could be monitored using a microwave monitoring system. The variation of the electromagnetic properties of the biological tissues and the magnetic implant during the therapy would result in a different propagation of the microwave signal. This work investigates the feasibility of using microwaves to non-invasively monitor hyperthermia treatments with a simplified monodimensional propagation model. The forward problem is solved to identify the working frequencies, the matching medium properties and study several candidate materials. By using the numerical solutions from nonlinear and multiphysics simulations of the bone tumor hyperthermia treatment using magnetic scaffolds, the microwave signal propagation dynamic is studied. From our feasibility analysis, we found that it is possible to correlate the average tumor temperature with significant ($\sim$20 dB) variations in the transmission coefficient during a typical interstitial hyperthermia session using magnetic scaffolds. Our work brings together, for the first time, the electromagnetic material properties, the physio-pathology and physics of the hyperthermia treatment and the microwave propagation problem, thus paving the route for the development of an innovative theranostic system.

In this letter, a microwave imaging strategy based on an artificial neural network (ANN) is applied, for the first time, to experimental data gathered from simplified neck phantoms. The ANN is used for solving the underlying inverse scattering problem, with the aim of retrieving the dielectric properties of the neck for monitoring and diagnostic purposes. The ANN is trained using simulated phantoms, to overcome the limited availability of experimental data. First, a simple configuration with a liquid-filled glass beaker is tested. Then, simplified 3-D-printed models of the human neck are considered. The preliminary findings indicate the possibility of training the network with numerical simulations and testing it against experimental measurements.

In the above article [1], several errors were introduced during the production.

Recounts the career and contributions of (Matteo Pastorino).

An inverse scattering method working at microwave frequencies for cervical diagnostics is proposed in this work. The aim is the diagnosis of cervical myelopathy, which is a disease that affects the first part of the spinal cord (between the C3 and C7 vertebra). A preliminary feasibility analysis oriented toward the development of an imaging system is reported. The system prototype includes a set of antennas that illuminate the neck and retrieve samples of the scattered electric field. The related inverse scattering problem is solved by using a nonlinear Newton-type reconstruction procedure, which provides two-dimensional images of the dielectric parameters of a neck cross section. A simplified cylindrical phantom mimicking the human neck has been designed for assessing the feasibility of the envisioned microwave measurement system and processing technique. Numerical results are reported to evaluate the capabilities of the proposed approach. Moreover, initial experimental results have been obtained by using cylindrical containers and a 3D printed version of the developed neck phantom.

Brain stroke treatments are strongly time-dependent. Indeed, the literature stresses the need for quick diagnosis to guarantee the most effective therapy with the famous motto “time is brain” [1]. Numerically speaking, a crucial goal for strokes is trying to achieve a door-to-needle time shorter than 60 minutes, called The Golden Hour [2], within which patients have the highest probability of a full recovery. This article presents a prototype of a Smart Helmet, a wearable device originally introduced in [3], [4], thought to be employed on site, to efficiently recognize and detect a brain stroke in a patient when the first aid medical team arrives. This work presents two different but complementary approaches: signal-processing-based and electromagnetic (EM). The former aims at outputting a two-level diagnosis: the presence (or not) of a stroke and (if it is present) of which type (ischemic or hemorrhagic), while the latter provides an accurate imaging of the brain that must be interpreted by medical staff. Both approaches are analyzed and compared, showing their strengths and highlighting their weaknesses.

An approach for monitoring the respiratory rate based on impedance measurements is presented in this paper. In order to possibly obtain a minimally-invasive wearable device, the electrodes are located on the head, near mastoid bones, and measure the variations induced in the electrical impedance by the physiological changes in the pharynx produced by respiration. The feasibility of the adopted configuration has been assessed by means of electromagnetic simulations involving a simplified model. Moreover, an ad-hoc data processing algorithm has been developed for extracting the breath rate from the measured signals. Preliminary experimental results are provided for investigating the capabilities of the developed setup.

Temperature monitoring during the cooldown in powder bed fusion of polymers is essential to qualify parts, and is an absolute novelty in the field. This is because the direct reading of the parts’ temperature is hindered by the powder. In this work, an innovative temperature monitoring technology based on microwave tomography is used to assess the part cooldown history in an industry-grade EOS P110 machine. So, relative and absolute readings of the temperature field can be carried out even for parts surrounded by powder. For the first time, several cooldown rates could be experimentally measured depending on the different polymer state: 0.56°Cmin−1in the liquid phase, 1.2°Cmin−1during supercooling, 0.83°Cmin−1in the solid phase. This technology will allow the development of a new generation of smart PBF machines that can better sense and hence control the entire PBF process, including the cooldown phase, which is often neglected but has the highest impact on the final mechanical performances of parts.

An electromagnetic inverse scattering procedure for the reconstruction of shallow buried objects in a homogeneous half-space is proposed. The approach is based on the numerical solution of the integral equations that model inverse scattering relationships, and it extends to strong scatterers the imaging capabilities of a previously proposed approach relying on approximated formulations. The inversion is based on the synergic application of a multifocusing strategy based on the iterative multiscaling approach along with an efficient regularization scheme based on the inexact Newton method. Numerical results corroborate the mathematical description to assess capabilities and current limitations of the proposed fully nonlinear technique.

In this paper the reconstruction of a shallow buried object is addressed by an electromagnetic inverse scattering method based on combining different imaging modalities. In particular, the proposed approach integrates the inexact Newton (IN) method with an iterative multiscaling approach. Moreover, the use of the second-order Born approximation is exploited. A numerical validation is provided concerning the potentialities arising by combining the regularization capabilities of the IN method and the effectiveness of the multifocusing strategy to mitigate the nonlinearity and ill-posedness of the inversion problem. Comparisons with the standard “bare” approach in terms of accuracy, robustness, noise levels, and computational efficiency are also included.

We introduce a new imaging technique that integrates the inexact Newton method into a multifocusing scheme within the contrast-source formulation of the inverse scattering problem. Representative results from an extensive validation concerned with both synthetic and experimental scattering data are reported to assess, also through comparisons, advantages and limitations of the proposed approach in terms of accuracy, robustness, and computational efficiency.

A recipe for Chocolate Hazelnut Ricotta Pie is presented.