In this paper, a novel framework is proposed which combines physical scattering models with artificial neural networks (ANN) to solve electromagnetic scattering problems of random media through a volume integral equation formulation. The framework is applied to a snow scattering problem where snow is represented by a bicontinuous random medium. A neural network is constructed linking the random media structure to the induced dipole moments on the media. The volume integral equation (VIE) serves as a natural physical constraint on the network input-out relations and is used to guide the training of the network. A discrete dipole approximation (DDA) strategy is adopted to convert the VIE into matrix equations which also defines the loss function of the surrogate neural network. For addressing deterministic scattering problems, this represents a viable alternative to traditional iterative algorithms, providing comparable accuracy at the expense of reduced efficiency. In solving statistical scattering problems, neural networks with physics-informed loss function achieve accuracy comparable to that of data-driven models while significantly reducing the dependency on extensive precomputed training datasets. The physics-based loss function also allows the network to self-diagnose the prediction accuracy in real operations. This work demonstrates a novel strategy to effectively merge physical equations with artificial neural networks, and the idea can be inspiring to many relevant fields, especially when randomness effects are exhibited through a complicated nonlinear system.

Nowadays, misalignment tolerant wireless power transfer systems providing simultaneous power supply to several devices are the subject of intensive research in the field of wireless charging of electronic devices. A critical parameter in such systems is the uniformity of magnetic field generated by a transmitting coil. In this paper, we examine the characteristics of the magnetic field distribution of arbitrary shape planar transmitting coils and propose a genetic algorithm for optimizing their design with the objective of increasing the field uniformity. This study stands out from existing literature by introducing an optimization approach that not only encompasses traditional circular and square coils but also extends to convex polygonal coils. The results of the algorithm are validated experimentally on coils of three various geometries including circular, square, and hexagonal coils. The coefficient of variation of the magnetic field, which serves as a quantitative measure of its uniformity, is found to be 3.6% for circular coil, 5.2% for square one, and 5.1% for hexagonal one in a region of interest encompassing half of the total area of transmitting coil.

The present study describes the architecture of a defected ground structure (DGS)-based eight-element multiple-input multiple-output (MIMO) antenna. It functions in the millimeter-wave spectrum n258 (24.25–27.5 GHz) and is mainly intendedd for fifth-generation (5G) applications. Each antenna element has an elliptical slot at the ground plane and a hook-shaped antenna integrated into it to reduce the mutual coupling among the adjacent antenna elements. With a wide bandwidth of 5 GHz and an isolation level greater than 37 dB, the proposed antenna effectively covers the frequency range of 22.5 to 27.5 GHz. At the designated resonant frequency, it generates a directed radiation pattern and gain of 6.31 dBi. The proposed antenna's diversity system performance characteristics are assessed using metrics like channel capacity loss (CCL), diversity gain (DG), mean effective gain (MEG), and envelope correlation coefficient (ECC). The 8-port proposed MIMO antenna prototype, exhibiting overall dimensions of 62.7 × 31 × 0.787 mm³, has been constructed utilizing an economical FR4 substrate, demonstrating a substantial association between the computed and measured outcomes, thereby establishing its viability as a potential candidate for 5G communications.

A uniform triangular array (UTA) is proposed for physics-based 2D direction-of-arrival (DOA) estimations of unknown incoming signals. Three capacitively loaded top-hat antennas are used as array elements. Unlike conventional array-based direction finding (DF) systems, complex antenna radiation patterns are used in array manifold calculations and DOA predictions, where coupling among array elements is naturally resolved. Both continuous wave (CW) signals and single tone signals with amplitude modulation (AM) are considered in DF simulations. Cramer-Rao bound (CRB) values are calculated to provide theoretical lower bounds of DF accuracies. Multiple signal classification (MUSIC) algorithm is used to further demonstrate the DF performance of the triangular antenna array without any angle ambiguities within the field of view from 0˚ to 360˚ in the azimuth direction and from 1.25˚ to 178.75˚ in the elevation direction.

This paper presents the design and in-depth performance analysis of a miniaturized four-port multiple-input multiple-output (MIMO) antenna module intended for integration on the rear cover of mobile devices. The four antenna elements are configured in a sequential rotation layout and fabricated on a low-profile circular substrate. Each antenna element features an E-shaped patch on the upper side of the substrate, coupled with a rectangular defected ground structure (DGS) on the lower side. A needle-like decoupling mechanism has been incorporated to improve the isolation between the antenna elements. The measured -10 dB impedance bandwidth ranges from 3.5 to 5.45 GHz, successfully meeting the demands of the 5G NR bands N77 (3.3-4.2 GHz), N78 (3.3-3.8 GHz), N79 (4.4-5 GHz), as well as the wireless local area network (WLAN) band (5.15-5.35 GHz). The isolation levels between the antenna elements exceed 17 dB. The average total efficiency is over 40.76%, and the envelope correlation coefficients (ECCs) are maintained below 0.01. The measurement outcomes indicate that the proposed MIMO antenna not only fulfills the requirements for the 5G and WLAN frequency bands but also successfully achieves miniaturization and superior wireless communication performance.

In this paper, a compact Dual-Band Bandpass Filter (BPF) has been proposed and designed based on two concepts. Firstly, transmission lines (TLs) feed the center resonators (CRs), which resonate at 11 GHz, while the metamaterial-based Center Resonators (CRs) resonate at 6 GHz for RF/Microwave applications. To miniaturize the planer structure, two different types of fractal geometry have been applied. First iteration modified-Minkowski fractal geometry has been applied on the CRs while the meander line with second order has been applied on the TLs. The proposed structure has been designed by using a Rogers RO4003 substrate with a thickness of 1.5 mm and a dielectric constant of 3.5. The simulation is implemented using CST microwave studio. To validate the proposed structure, the compact dual-band BPF is fabricated, and the measurements show high agreement with the simulation. Finally, the proposed structure achieves a 26 % reduction compared to the previous work.

This paper introduces a novel finite element boundary integral approach for magnetic resonance electrical properties tomography (MR-EPT) with an inhomogeneous background which improves imaging quality by utilizing inhomogeneous background inversion and allows for a flexible selection of areas for fine reconstruction, thereby saving resources and quickly obtaining the most important information. In the proposed approach, a fictitious inhomogeneous background is initialized, followed by a preliminary reconstruction conducted across the entire field of view (FOV) through a few iterations. This fictitious inhomogeneous background aims to enhance the quality of reconstruction, surpassing that achieved through inversion in a homogeneous background. The proposed method is significantly suitable to the prevailing refinement mechanism, where the refinement area identified from the preliminary reconstruction image is embedded in an inhomogeneous background. This method combines the advantages of the computational efficiency of local methods and the noise robustness of global methods. Numerical examples have validated that the inversion with a fictitious inhomogeneous background yields a superior reconstruction quality. The subsequent narrowing of the inversion area results in a more focused inversion process, significantly reducing reconstruction time.

This paper presents a four-port quarter-mode substrate integrated waveguide (QMSIW) MIMO antenna designed for 2.1 GHz wireless applications. The antenna employs orthogonally positioned complementary square-split ring resonator slots to achieve substantial miniaturization. Additionally, mutual coupling between the antenna elements is effectively minimized by incorporating cross-shaped slots between them, enhancing overall performance. The proposed four-port MIMO antenna achieves high isolation of 40 dB and features a compact electrical size of 0.19λ₀ × 0.19λ₀. The antenna demonstrates outstanding MIMO performance, with simulated and measured gains of 5.32 dBi and 5.44 dBi, respectively. Its efficiency is further supported by key performance metrics, including a low envelope correlation coefficient (ECC) of 0.0841 and a high diversity gain (DG) of 9.22 dB, ensuring enhanced signal reliability and reduced interference. With its compact structure, excellent isolation, and strong diversity performance, the proposed antenna serves as a highly suitable candidate for directional Wi-Fi applications.

This work presents a quintuple-mode wideband bandpass filter utilizing a cylinder cavity loaded with a metallic perturbation on each base and a gap between them. The cylindrical cavity is fed with two inline disk-loaded feeding lines placed in the middle of the cavity wall. Two inline shorting pins are placed orthogonally in the same feeding lines plane. The five resonant modes are: TM₀₁₁, TE₁₁₁, the degenerate modes of TE₂₁₁ and quasi-TE₃₁₁ excited by the pair of disk-loaded feeding lines and shorting pins. This combination allows the generation of two transmission zeros (TZs) out of the passband. The mode analysis and geometric configuration are well studied and presented in this work. The results of the quintuple-mode filter demonstrated a fractional bandwidth of 66.87% at a central frequency of 3.69 GHz with insertion loss 0.18 dB and return loss higher than 19.4 dB. The coupling matrix of the proposed filter is synthesized as a five-pole Chebyshev filter showing close alignment with the simulated results. Finally, the proposed bandpass filter is fabricated and measured for comparison. Close agreement is achieved between the simulated and measured results.

Due to the rapid development of the internet and mobile communication needs, visible light communication (VLC) has become an attractive technique for indoor wireless communication. This research investigates how data rates in VLC systems are affected by the wavelength division multiplexing (WDM) technology. The WDM technique allows different data streams to be transmitted simultaneously over different wavelengths in the same optical channel. In WDM-VLC systems, the interference between channels known as crosstalk is a significant problem that may reduce the quality of communication. By optimizing the field of view (FOV) of the optical receiver and changing system parameters to reduce interference, this research resolves the crosstalk issue. With attention to the signal-to-interference-plus-noise ratio (SINR) and channel bandwidth, we use a simulated indoor environment to examine how line of sight (LOS) and non-line of sight (NLOS) elements affect the system performance. The results show that reducing the FOV leads to reducing the crosstalk and significantly enhancing data speeds and reliability in the system. Additionally, a review of practical challenges related to the implementation of different FOV lenses is presented, along with a comparative assessment of complexity, scalability, and cost in relation to present solutions. The results offer important updated knowledge on WDM's capabilities in VLC systems, enabling rapid transfer of data and efficient lighting for smart interior spaces.

As a new type of motor, axial flux permanent magnet synchronous motor has the advantages of compact structure and high power density. It shows good application prospects in new energy vehicles, unmanned aerial vehicles, and other fields. However, axial flux permanent magnet synchronous motor needs to consider the balance of multiple objectives during the design process, which makes its optimal design a complex multi-objective optimization problem. Therefore, the study proposes a motor optimization method based on multi-objective genetic algorithm. The method optimizes the rotor and stator parameters of the motor by establishing an analytical model of the motor's magnetic field and combining it with a multi-objective genetic algorithm. The experimental results indicated that the optimized motor with multi-objective genetic algorithm reached 92% in terms of efficiency, 25% in terms of power density; energy consumption was reduced to 2.5 kWh; failure rate was reduced to 1.5%; and noise level was reduced to 65 dB. In addition, the multi-objective genetic algorithm significantly improved the control stability index, which increased to 98%, indicating a more stable motor response under varying loads. The disturbance rejection capability was enhanced to 99%, demonstrating strong resistance to external noise and parameter fluctuations. Furthermore, the system response frequency reached 100 Hz, reflecting a faster dynamic response to input variations. It is indicated that the optimization method based on multi-objective genetic algorithm can effectively enhance the comprehensive performance of axial flux permanent magnet synchronous motor and significantly improve its competitiveness in high power density and high efficiency applications.

This paper investigates the effects of the step-skew rotor on electromagnetic vibration performance of a 72-slot/12-pole interior permanent magnet synchronous machine (IPMSM) considering the carrier sideband current harmonics. Firstly, the effect of the step-skew rotor on carrier sideband current harmonics generated by the space vector pulse-width modulation technique is investigated. Second, the interaction electromagnetic field near the first carrier frequency of the 72-slot/12-pole IPMSM is analytically investigated. The effects of the step-skew rotor on the electromagnetic force at low-medium and carrier frequency domain are discussed. The force modulation effect is also considered to investigate the effects of the step-skew rotor on the 72nd-order force and the zeroth mode vibration. Finally, the vibration responses of two IPMSMs with and without the step-skew rotor are simulated to validate the suppression effect of this structure on the 72nd-order force. The simulation results demonstrate that the step-skew rotor design effectively reduces vibration acceleration by 75.7% at 12 times of the fundamental frequency and 30.4% at carrier sideband frequency.

Aspiration pneumonia is a type of lung infection caused by the accidental inhalation of foreign substances into the respiratory tract. It is commonly seen in the elderly, young children, and individuals who are unconscious or have difficulty swallowing. Early detection and diagnosis of aspiration pneumonia are beneficial for improving patient outcomes and reducing the medical burden. In this study, we collected bronchoscopic video data from 25 patients in two hospitals. After image preprocessing and expert annotation, we obtained 2830 images from some patients for training and 1215 images from the other patients for validation. We selected three deep learning methods for training. The experimental test results for the identification of aspiration pneumonia showed that ResNet-50, which is based on convolutional operations, gave the best performance in the automatic identification of aspiration pneumonia, with a precision of 97.82%, a recall of 91.82%, an F1 score of 94.73%, and an overall accuracy of 95.88%. The experiments demonstrated that deep learning methods can be used for the automatic identification and diagnosis of aspiration pneumonia from bronchoscope images and deep learning is reported here for the first time for diagnosing aspiration pneumonia from bronchoscope images.

Numerical modeling of eddy current (EC) phenomena is pivotal in nondestructive evaluation (NDE). It has become invaluable in NDE industries, contributing to probe design, inspection procedures, defect characterization, model training, and results interpretation. This study comprehensively explores two numerical methods - Volume Integral Method (VIM) and Finite Element Method (FEM) to assess their suitability for EC NDE. Four test cases involving varying geometries, defect types, and probe configurations were modeled to compare computational compatibility. Numerical results are evaluated for their accuracy, efficiency, and practical implications. Results indicate a reasonable correlation between the two methods, with VIM excelling at computational efficiency for simpler geometries, and FEM demonstrating robustness for complex configurations. The findings highlight the strengths and limitations of each method, aiding users in selecting appropriate techniques for defect characterization and optimizing inspection conditions.

A dielectric resonator antenna (DRA) with the resonator body formed, or compounded, by multiple building blocks is proposed. The approach gives flexibility in adjusting the resonator's shape to control input impedance and resonance frequency of the antenna. A simplified method of attaching the resonator's building blocks to the grounded dielectric substrate allowed for reduced fabrication complexity and manual reconfigurability of this compound DRA (cDRA). Several cDRAs with variable resonator sizes were studied theoretically and experimentally.

Metamaterial-based absorbers offering perfect and broadband absorption are greatly desirable in radar and stealth technology systems. Further, the polarization insensitive feature makes the absorber useful for industrial applications. This study examines the design of a wideband metamaterial-based radar absorber functioning throughout the frequency range of 8 to 16 GHz. The metasurface design comprises two concentric metallic rings and a fan-shaped circular metallic disc at the middle. The rings and the central disc are interconnected with surface-mounted chip resistors to enhance the absorption bandwidth. It is depicted that more than 90% absorptivity was attained from 8 to 16 GHz. A wideband absorber with angular stability and polarization insensitivity is an excellent choice for wireless communications, particularly in radar applications. Further, the measured results of the prototype corresponded well with the simulated outcomes.

Antenna gain is an important metric for most modern communication systems. The most common method for determining antenna gain is to produce an incident plane wave using a reference antenna and measure the received power at the antenna of interest. By measuring the received power several wavelengths away in an isolated environment, such as an anechoic chamber, the gain of an electrically small antenna, dimension less than one wavelength, can be determined. This and other similar methods work well for frequencies above 2 GHz, but lower frequency measurements can be logistically challenging and expensive due to the large facilities required and the lack of readily available broadband absorber materials. This work presents a new method for determining antenna gain using a transmission-line-based near-field S-parameter measurement in a waveguide. To provide evidence for the proposed method, two monopole antennas are modeled over an infinite ground plane using full-wave electromagnetics, and both are experimentally measured within a waveguide. Good agreement was found between the model and measurement, providing evidence of the validity of the method.

This paper proposes a broadband dual-polarized (DP) magneto-electric dipole antenna (MEDA) loaded with parasitic elements. The antenna structure comprises a rectangular reflecting cavity, four corner-truncated and optimized horizontal electric dipoles, eight vertically arranged magnetic dipoles, and four triangular parasitic elements embedded in the truncated corners of the electric dipoles. An orthogonal Γ-shaped feeding network is employed to excite dual-polarized modes, ensuring stable bandwidth and gain performance. Experimental measurements demonstrate an overlapping impedance bandwidth (|S₁₁| and |S₂₂| ≤ -10 dB) of 84.7% (1.92-4.74 GHz) for both ports, with an in-band peak gain of 11.69 dBi and port isolation exceeding 27 dB. This design provides a compact high-isolation dual-polarized antenna solution for 5G multi-band communication systems.

To address the quantification challenges of transient responses in direct lightning strike protection design for photovoltaic (PV) power generation systems, this study establishes an integrated coupled model using CDEGS simulation software, incorporating horizontally layered soil, PV mounting structures, and grounding systems. A comprehensive consideration of key factors, including lightning current waveforms, soil resistivity, and the number of down conductors and vertical grounding electrodes, enables quantitative analysis of transient overvoltage and transient ground potential rise (TGPR) distribution characteristics under varying operating conditions. The results demonstrate that both soil resistivity and lightning current waveforms are critical factors influencing the transient lightning-induced characteristics of PV systems. In typical low-resistivity (ρ = 200 Ω.m) and high-resistivity (ρ = 2000 Ω.m) soil environments, increasing the number of grounding down conductors and vertical grounding electrodes can both reduce induced overvoltage and transient ground potential rise. However, beyond a certain threshold, shielding effects between adjacent grounding bodies limit current dissipation efficiency, leading to diminishing returns. Therefore, PV system lightning protection design must holistically account for soil properties, lightning current parameters, and optimized layout strategies to mitigate transient amplitudes, achieving an optimal balance between lightning protection effectiveness and economic efficiency.

This article summarizes the author's work over the years having the goal of developing computational tools for deter-mining the quantitative distribution of where radiation is emitted from objects of interest.