Search

A highly sensitive, large throughput and wide dynamic range (in the aspect of size and material type) broadband tunable microwave interferometer based rectangular waveguide wear debris detection system is presented. Passing of (non)ferrous and (non)conductive particles through the sensor, which causes inductance and/or capacitance changes, are detected as transmission signal variations of the system. We first demonstrate the sensing principle with the aluminum particle of 800 μm and epoxy resin particle of 1 mm3 at 6 GHz. Then, the effects of particle size and flow speed are tested with aluminum particles ranging from 200 μm to 1.8 mm. The measurement results show that the system is able to measure particle sizes with known material and differentiate nonconductive particles from conductive ones. With the proposed data processing techniques, this system is capable of detecting as small as 76 μm particles within an 8 mm outer diameter tube. And with future improvements on sensitivity and particle material characterization, the proposed system has the potential to differentiate (non)ferrous and (non)conductive particles and measure their sizes as a real-time wear debris sensing devices for rotating and reciprocating machines.

This article presents the design of integrated antennas in the $D$ -band (110–170 GHz). First, an electromagnetic analysis is used to tune the propagation characteristics of surface waves supported on an artificial magnetic conductor (AMC). Next, a subarray of two miniaturized dipoles is designed where each dipole attains axial resonance and the separation between them is optimized for substrate wave cancellation. This subarray simultaneously performs spatial power combining and substrate wave cancellation, leading to significant increase of the effective isotropic radiated power (EIRP). Two miniaturized dipoles with axial length of $252~\mu \text{m}$ and spacing of $450~\mu \text{m}$ was fabricated in 28 nm bulk CMOS with low resistivity silicon ( $10~\Omega \cdot $ cm). Each dipole is fed by a cascade of two triplers and a power amplifier. Measurements show a gain of 2.3 dBi, EIRP of 12.2 dBm at 139 GHz and a 3 dB bandwidth of 127–154 GHz (19% fractional) without using any additional substrates, lenses, or postprocessing steps. Scalability of this technique for forming larger arrays is also presented.

In this article, a new wideband cavity-backed bow-tie microstrip antenna is presented for short-range 77–81 GHz radar applications. The design is based on a subarray consisting of two single elements. Each element consists of a multilayer bow-tie patch antenna surrounded by a metallic cavity to suppress the surface waves and parallel plate modes. The subarray elements are fed through a laminated waveguide with 180° phase shift. An eight-layer build-up and a new advanced multilayer PCB technology, the so-called any-layer PCB technology, is used for the fabrication. The performance of this antenna has been verified by both simulations and measurements. A bandwidth of more than 8% is achieved. The gain is 7.9 dBi, and the radiation efficiency is 70% at 79 GHz. Beam widths of 84° in the E-plane and 48° in the H-plane are obtained, which are compatible with radar requirements for short-range applications.

IEEE A simplified dielectric material characterization technique that combines an equivalent impedance algorithm and a general “line–line” trace method is introduced. The technique is applicable to every type of transmission lines. Due to the convenient fabrication and the allowance of integration with various polymers, coplanar waveguide transmission lines integrated with an SU-8 microfluidic channel is primarily used in this study. The conformal mapping method, an efficient and straightforward way, is introduced to build the foundation of dielectric material characterization and to optimize the sensor design. A validation measurement of the proposed technique is performed on deionized water, and show valuable consistency with previous reliable data presented in the literature. Next, the technique is applied on solid SU-8 characterization with four groups of on-wafer measurements. SU-8 measurement results and the related uncertainty analysis are demonstrated subsequently. ispartof: IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY vol:61 issue:5 pages:1639-1646 status: published

Due to its noninvasive property, the interdigital capacitor (IDC) has been applied in dielectric liquid detection and characterization. In order to integrate the IDC sensor on a lab-on-chip, it is often required to minimize and optimize the sensor for sensitive and efficient performance. However, the conventional numerical simulation approach is quite time-consuming. Therefore, an efficient analytical method is proposed herein, leading to accurate capacitance and conductance expressions of an arbitrary multilayer-structured IDC. The model is validated with practical measurements of a series of coplanar waveguide (CPW) structure-based IDCs. In addition, an accurate characterization function, which relates the IDC capacitance and conductance to the complex permittivity of a material loaded on the top of the IDC sensing area, is obtained. The characterization function shows good agreement with the finite-element method (FEM) simulation results, which demonstrates the capability of the IDC sensor in dielectric spectroscopy measurements of $\mu \text{L}$ and even nL liquids.

In this article, a miniaturized microwave resonator is designed and fabricated. A loop-mode resonance is created when the permittivity difference between the two materials loaded in the two fluidic channels is sufficiently large. A highly concentrated electrical field is generated in a microfluidic channel loaded with deionized water to achieve high sensing spatial resolution within a sensing volume of 45 pL. On-wafer measurements are performed by keeping one channel unloaded as a reference while loading the other channel with different mixtures of isopropanol and deionized water. The acquired microwave-microfluidic measurements confirm that the proposed resonator can detect changes in the permittivity of the loaded liquid between the values of 3.2 and 42.4, making it a promising low-cost and label-free approach in applications requiring novel sensing techniques such as mixing monitoring and flow cytometry.

Drones are currently being explored for safety-critical applications where human agents are expected to evolve in their vicinity. In such applications, robust people avoidance must be provided by fusing a number of sensing modalities in order to avoid collisions. Currently however, people detection systems used on drones are solely based on standard cameras besides an emerging number of works discussing the fusion of imaging and event-based cameras. On the other hand, radar-based systems provide up-most robustness towards environmental conditions but do not provide complete information on their own and have mainly been investigated in automotive contexts, not for drones. In order to enable the fusion of radars with both event-based and standard cameras, we present KUL-UAVSAFE, a first-of-its-kind dataset for the study of safety-critical people detection by drones. In addition, we propose a baseline CNN architecture with cross-fusion highways and introduce a curriculum learning strategy for multi-modal data termed SAUL, which greatly enhances the robustness of the system towards hard RGB failures and provides a significant gain of 15% in peak $\boldsymbol{F_1}$ score compared to the use of BlackIn , previously proposed for cross-fusion networks. We demonstrate the real-time performance and feasibility of the approach by implementing the system in an edge-computing unit. We release our dataset and additional material in the project home page.

Event-based neural networks are currently being explored as efficient solutions for performing AI tasks at the extreme edge. To fully exploit their potential, event-based neural networks coupled to adequate preprocessing must be investigated. Within this context, we demonstrate a 4-b-weight spiking neural network (SNN) for radar gesture recognition, achieving a state-of-the-art 93% accuracy within only four processing time steps while using only one convolutional layer and two fully connected layers. This solution consumes very little energy and area if implemented in event-based hardware, which makes it suited for embedded extreme-edge applications. In addition, we demonstrate the importance of signal preprocessing for achieving this high recognition accuracy in SNNs compared to deep neural networks (DNNs) with the same network topology and training strategy. We show that efficient preprocessing prior to the neural network is drastically more important for SNNs compared to DNNs. We also demonstrate, for the first time, that the preprocessing parameters can affect SNNs and DNNs in antagonistic ways, prohibiting the generalization of conclusions drawn from DNN design to SNNs. We demonstrate our findings by comparing the gesture recognition accuracy achieved with our SNN to a DNN with the same architecture and similar training. Unlike previously proposed neural networks for radar processing, this work enables ultralow-power radar-based gesture recognition for extreme-edge devices. ispartof: IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS vol:34 issue:6 pages:2869-2881 ispartof: location:United States status: published

Uncertainty analysis is performed on a coaxial probe-based dielectric spectroscopy measurement platform, and its impact on the modeling of dielectric relaxation in an aqueous solution is further analyzed. Based on the characterization theory and calibration principle of the platform, uncertainty due to the different available expressions and parameters of the calibration liquid is calculated. Then, repeatability uncertainty related to drift of the measurement platform, environment change, position variation of the probe immersed in the liquid under test (LUT), and so on is analyzed. Next, three types of water–alcohol mixtures and one group of glucose solutions are measured using the platform. Based on the measurement results, the best model is chosen to analyze the frequency dependence of the solutions’ relaxation properties. In addition, based on the uncertainty analysis, appropriate mathematical formulas are provided to model the composition dependence of the solutions’ relaxation parameters, which can be used as a liquid detection technique for chemical and biological applications.

In this paper, a novel microwave sensor is realized by double stub matching a deionized (DI) water loaded interdigital capacitor to 50 Ω to create a minimum in reflection coefficient around 1.5 GHz. Glucose, l -proline and glycine solutions with different concentrations are loaded into the sensor and consequently impact the frequency of the minimum reflection coefficient. The proposed sensor achieves a linear correlation between its minimum reflection coefficient frequency and permittivity of the loaded solutions. Mixtures between fetal bovine serum (FBS) and phosphate-buffered saline (PBS) having FBS volume fractions of 0%, 20 %, 40 %, 60 %, 80 % and 100 % are also loaded into the proposed sensor. A linear relationship between the volume fractions of the loaded mixtures and the minimum reflection coefficient frequency point of the proposed sensor is as well achieved. Measurement results demonstrate the potential of the proposed miniaturized sensor, coupled with the dislocated-of-the-chip and space-consuming supporting microwave circuitry, for applications such as sample preparation at the microfluidic scale.

A nondestructive complementary split ring resonator (CSRR), working in the 0.5 to 6 GHz frequency range, is presented in this work. Three CSRR designs, on the same 5 mm by 5 mm footprint, with different internal dimensions were designed, fabricated, and measured. The results demonstrate a large shift in resonant frequency, up to 2.544 GHz, over a broad relative permittivity range from 1 to 80, and a large penetration depth, up to 7 mm, into the material under test. Secondly, this work presents a deconvolution-based method to increase the spatial resolution of the presented sensors compared to other sensing structures and algorithms. A conformal mapping function relating the resonant frequency to the relative permittivity was fitted through measurements and the point spread function (PSF) was derived from a known dielectric permittivity distribution. Experimental results demonstrated that the extracted PSF was able to reconstruct the dielectric distribution in six different scenarios. Additionally, a spatial resolution of 2 mm has been established according to the Rayleigh criterion, and objects smaller than the size of the sensor have been reconstructed as well. Objects with a different permittivity values can be distinguished as well. The reconstructed images of the objects have a good signal-to-clutter ratio up to 12.0%, which results from a smooth recovered background, and no spatial offset was present between the original and reconstructed objects.

ispartof: INTERNATIONAL JOURNAL OF NUMERICAL MODELLING-ELECTRONIC NETWORKS DEVICES AND FIELDS vol:34 issue:1 status: Published online

The lack of a standardized operating protocol for dielectric measurements of biological tissues makes it difficult to objectively compare reported dielectric properties since these properties are affected by multiple environmental confounders. This work investigates the effect of dehydration on dielectric properties as a function of time at different temperatures. Additionally, it proposes a dehydration mitigating technique. The dielectric properties of porcine liver samples were measured during a 35 minute measurement interval at different temperatures. A first set of six experiments was conducted in standard laboratory conditions. A second set of two experiments was conducted with a modified measurement setup with reduced air flow. Under standard conditions, relative changes in the dielectric properties up to 9% were observed. With the modified setup, the relative change was halved compared to standard conditions. This reduction indicates that the environmental conditions have a considerable influence and that modifying these conditions extends the stability of the dielectric properties over time. For several applications relying on the difference in dielectric properties between healthy and malignant tissue, an error of 9% in measured data is on the same order as the effects of interest. The expected difference in dielectric contrast could be obscured due to dehydration.

A microwave microfluidic sensor for detecting binary liquid mixtures with a dielectric method at RF/microwave frequencies is presented in this article. The sensor is based on a split-ring resonator (SRR) that is implemented in a microstrip transmission line, with interdigitated electrodes (IDEs) being integrated into the ring for liquid detection. Based on the equivalent circuit of the IDE-SRR device and with a series of finite element simulations, the detection theory is developed, and the device design optimization is investigated. The validation measurements on water–isopropanol liquid mixtures with various concentrations show that the proposed IDE-SRR sensor has higher sensitivity than the previous standard SRR sensor. The IDE-SRR sensor is then used to detect two binary liquids, i.e., water–methanol mixtures and water–tetrahydrofuran mixtures. The measured effective permittivity results of the binary mixtures at RF/microwave frequency range are compared with the existing mixing models for binary dielectric mixtures at zero frequency.