The Fourier Neural Operator (FNO), a recently proposed neural network designed for solving partial differential equations (PDEs), is being explored to accelerate electromagnetic (EM) simulations. FNOs are data-driven models that learn to map material distributions to electromagnetic field solutions based on a full-wave simulation training dataset. Unlike conventional computational electromagnetic solvers that handle individual instances of Maxwell’s equations, FNOs learn an entire family of these equations. The Fourier Neural Operator excels in global feature extraction. When configured with cascaded multilayer Fourier Neural Operators, the network functions similarly to iterative solvers like the Born series expansion used for solving integral equations. In our electrostatic study, we train an FNO to predict electric potential given an arbitrary permittivity distribution in a unit square. In the electrodynamic study, we apply the FNO to an arbitrary-shaped scattering problem involving 2 D TMz coupled Maxwell’s equations. Numerical and electromagnetic benchmarks show that the FNO achieves state-of-the-art accuracy and can be up to three orders of magnitude faster than traditional finite difference method (FDM) solvers.

The demand for lightweight antennas in 5 G/6 G communication, wearables, and aerospace applications is rapidly growing. However, standard manufacturing techniques are limited in structural complexity and easy integration of multiple material classes. Here we introduce charge programmed multi-material additive manufacturing platform, offering unparalleled flexibility in antenna design and the capability for rapid printing of intricate antenna structures that are unprecedented or necessitate a series of fabrication routes. Demonstrating its potential, we present a transmitarray antenna composed of an interconnected, multi-layered array of dielectric/conductive S-ring unit cells, reducing 94% mass of conventional antenna configurations. A fully printed circular polarized transmitarray system fed by a source and a Risley prism antenna system operating at 19 GHz both show close alignment between testing results and numerical simulations. This printing method establishes a universal platform, propelling discovery of new antenna designs and enabling data-driven design and optimizations where rapid production of antenna designs is crucial.

This communication presents a novel adaptive solution-space methodology integrated with particle swarm optimization (PSO) to dynamically adjust search boundaries in response to swarm movement within the search space. Traditional PSO can be limited when the global optimum lies outside the initially defined search space, leading to extended search times without reaching target fitness values. The proposed adaptive approach overcomes this limitation by automatically expanding the solution space in relevant dimensions, promoting efficient convergence toward the global optimum while minimizing manual intervention. The performance of the proposed method is assessed through tests on mathematical benchmark functions and three different antenna designs: a dual-band microstrip patch antenna, a six-element Yagi-Uda antenna, and a six-element series-fed microstrip patch antenna. The dual-band and series-fed antennas were simulated in Ansys HFSS using a custom Python-HFSS interface, while the six-element Yagi-Uda antenna was optimized using MATLAB. Results indicate that the adaptive solution space significantly enhances PSO's effectiveness, achieving optimal design configurations across diverse electromagnetic applications, demonstrating reduced computational costs and improved optimization accuracy. The methodology described in this communication could also be applied to other nature-inspired optimization techniques.

Unmanned aerial vehicles (UAVs) have gained attention in antenna near-field measurements for outdoor large antennas due to their ability to reduce the amount of required test equipment. The most promising UAV-based setup for these outdoor measurements is an amplitude-only approach as amplitude-only measurements are robust against positioning errors which is an inherent problem facing UAVs. Current implementations of UAV amplitude-only near-field scans utilize matrix solvers and equivalent currents to account for the non-uniform data collection when performing phase reconstruction. In this paper, the conjugate gradient non-uniform fast Fourier transform (CG-NUFFT) algorithm is utilized for phase reconstruction from non-uniform two planar amplitude-only measurements. The system is tested through simulations of a dipole-element phased array antenna with various tapering and steering angles, aiming to reconstruct the far-field patterns. UAV non-uniform positioning is also simulated by introducing deviations to the targeted uniform waypoints. Simulation results demonstrate that the CG-NUFFT algorithm accurately reconstructs far-field patterns.

The growing need for compact, high-gain antennas in space communication and remote sensing radar systems, particularly for CubeSat integration, has led to the development of a compact and lightweight transmitarray antenna (TA). Our novel design uses two layers of ultra-thin (0.5 mil) Kapton membrane, allowing for both low mass and easy deployment. The S-shaped metasurface geometry enables asymmetric transmission and facilitates 90° linear polarization conversion while offering full $2\mathfrak{n}$ phase coverage and maintaining high transmission efficiency for linearly polarized feed source. Simulation results indicate that the 10 cm diameter TA design can achieve 29.6 dBi directivity and 63% aperture efficiency at 35.75 GHz, showing great potential for future millimeter-wave space communication applications.

In this article, we introduce and investigate a hybridization algorithm based on particle swarm optimization (PSO) and brainstorm optimization (BSO). The hybrid BSO–PSO (HBPSO) technique adopts PSO that is initialized by BSO within the starting iterations. The performance of HBPSO is significantly enhanced compared to single BSO or PSO when applied to high-dimensional optimization problems with local minima. The hybrid procedure is validated by showing appropriate convergence curves when applied to six benchmark functions. Guidelines regarding the selection of the inertial factor and switching iteration are investigated and presented accordingly. The proposed HBPSO is then validated using practical optimization tasks. It is demonstrated that HBPSO can outperform single PSO or BSO techniques in addressing representative antenna-related problems, including patch antenna circuit model extraction, conformal antenna array synthesis, and full-wave antenna design problems.

A key design metric for near-field focused antennas is its resolution, which is the spatial extent over which the fields decay by 3 dB. While the resolution for far-field focused antennas is widely appreciated, a detailed understanding of the resolution of near-field focused antennas has not been presented. A unique requirement of near-field focused antennas is the requirement to characterize both transverse and axial resolutions. We begin by revisiting the classical scalar Fresnel-Kirchhoff's diffraction integral and the assumptions involved in obtaining a closed-form solution of the integral. These closed-form solutions are then analyzed in depth to reveal the dependency of the resolution on the array size, focal length, and the frequency (wavelength) of operation. These observations are validated via a representative phased array consisting of infinitesimal dipoles which radiates vector fields. This simulation model serves as the basis to assess the domain of applicability of the closed-form solutions and reveals insightful dependencies of the near-field resolution on the geometry and frequency. This model is further employed to study the impact of focal-spot scanning on the resolution of near-field focused antennas.

In one's life, there are certain individuals who serve as constant pillars of support and admiration. Ross, unquestionably, belonged to that esteemed group. His extensive knowledge across a multitude of subjects and his unwavering commitment to active engagement and selfless contributions exceeded all expectations. Within our IEEE AP-S and USNC-URSI organizations, he truly embodied the role of an encyclopedia. It is with great humility that I present this invited talk during a special session at the 2024 75th Anniversary of the IEEE AP-S in Florence, Italy, dedicated to the professional legacy of the late Ross Stone (1947–2023). My association with Ross spans nearly five decades, during which we shared numerous fruitful technical and professional exchanges on various electromagnetic topics, such as inverse scattering and remote sensing. Our initial encounter took place at the 1973 IEEE AP-S/USNC-URSI conference in Boulder, Colorado, where I first met Ross and presented my first conference paper. What remains vivid in my memory is Ross's avid interest in overseeing edited publications, bringing together exceptional contributors, and producing top-notch special issues encompassing diverse subjects. Interestingly, our paths converged through extensive participation in national and international conferences, collaborative contributions to committee meetings, and a myriad of shared experiences.

Throughout human history, the reflector antenna has found diverse applications due to its ability, among various antenna configurations, to offer the highest gain, widest bandwidth, and superior angular resolutions at relatively low costs. In simple terms, the primary function of a reflector antenna is to concentrate or emit a significant portion of electromagnetic energy across its aperture into a focal plane during reception or to radiate towards far fields for communication or energy transfer during transmission. Conventional reflector antennas utilize conic sections such as the parabola, ellipse, hyperbola, and sphere to focus or efficiently radiate electromagnetic waves. Reflector antennas are typically classified based on radiation pattern type, reflector surface type, and feed type. Pencil-beam reflectors, known for their maximum gain and fixed beam directions upon installation, are particularly popular in point-to-point microwave communications and telemetry. In satellite communication systems, the uplink pencil-beam is often steered by moving the reflector or adjusting it within a limited range using the feed. Recent iterations of satellite reflectors have introduced other prevalent radiation pattern classifications, including contour (shaped) beams and multiple beams. These applications necessitate reflectors with enhanced off-axis beam characteristics and non-standard conical shapes. The demand for high-performance large reflector antennas in space applications has driven the development of various deployable concepts, such as mesh and inflatable designs. Furthermore, the realms of radio astronomy and deep space communications have spurred captivating developments and engineering advancements in reflector antenna technology.

This article provides a glimpse into my career journey over the years, highlighting the valuable lessons I’ve learned. It summarizes 15 key career life lessons that I encourage our younger colleagues to embrace. By putting these lessons into practice, positive outcomes are sure to follow.

There are certain people in one's life that you always rely on and admire. Ross was for sure one of them. His knowledge of so many diverse topics and his abounded dedication to stay engaged and contribute so unselfishly were beyond anyone's expectation. He was for sure the encyclopaedia of our USNC-URSI and IEEE AP-S organizations. I am humbled and delighted to present this invited talk in this special session at the 2024 50 th Anniversary of USNC-URSI Boulder Colorado Meeting dedicated to the professional life of the late Ross Stone (1948–2023). I knew Ross for nearly five decades and had many fruitful technical and professional interactions on all kinds of EM topics including inverse scattering, non-uniqueness, remote sensing, etc. As a matter of fact, the 1973 USNC- URSI/IEEE AP-S conference held in Boulder, Colorado, was our first conference where I delivered my first paper. What I clearly remember is Ross' keen interest in managing edited publications by bringing outstanding contributors together and producing first rate special issues on diverse topics with historical content. Ironically we shared some similar experiences by participating in enormous numbers of national and international conferences and contributing to many committee meetings together and the list goes on and on. We all lost an exceptional editor, technical contributor and a dear friend.

This paper introduces an innovative adaptive solution methodology that integrates with the Particle Swarm Optimization (PSO) algorithm. The methodology is specifically designed to autonomously adjust the solution space, particularly in dimensions crucial to the design space. In situations where the global optimum is beyond the initially defined solution space, the particles may struggle to attain the desired fitness value despite extensive searching. To overcome this challenge, the proposed methodology enables the optimization algorithm to operate efficiently, even when the global optimum lies outside the initial bounds. The performance of the adaptive solution space methodology undergoes evaluation using both mathematical benchmark functions and an antenna optimization application. Additionally, the paper conducts a comparative analysis with a method involving the expansion of the solution space in all dimensions after a specified iteration. The results not only showcase the effectiveness of the proposed method in dynamically expanding the solution space but also demonstrate a significant improvement in convergence speed compared to an unspecified expansion method.

We introduce an optimization assisted by a neural network (ONN) predictor to the electromagnetic community. ONN belongs to the class of the surrogate model-based optimization approaches, and approximates the objective function using a non-linear approximator – neural network. We provide a comprehensive description of the ONN algorithm and the details of its mathematical formulations. We apply ONN to optimize three popular benchmark functions and compare its performance with some commonly used optimization algorithms, namely particle swarm optimization, genetic algorithm, and Bayesian optimization. For the first time, we demonstrate ONN’s applicability and effectiveness in antenna design problems by optimizing a six-element Yagi-Uda antenna and by solving a challenging ten-dimensional dual-band slotted patch antenna constrained optimization problem. To achieve this, ONN is linked with a full-wave electromagnetic simulation solver through an application user interface. The optimized slotted patch is fabricated and measured to demonstrate how ONN can be part of the full antenna design process. Our empirical results indicate that ONN requires less objective function evaluations to reach the same qualitative point and reaches better optimal points for the same number of iterations for the studied benchmark functions and antenna optimization problems compared to the aforementioned baseline optimization algorithms.

In the past decade, CubeSats have undergone a revolution, moving from university research projects to enabling industry opportunities and government missions. In 2014, the Jet Propulsion Laboratory, California Institute of Technology (JPL/Caltech) initiated a research and technology development effort to advance CubeSat communication capabilities. One of the critical thrusts was the Ka-band parabolic deployable antenna (KaPDA). This antenna started with the ambitious goal of fitting a 42 dB, 0.5 m, 35 GHz antenna in a 1.5U canister. At that time, there had been minimal development in high-gain CubeSat antennas, critical for high-data-rate communications and remote sensing science. A Ka-band high-gain antenna would provide a 10,000 times increase in data communication rates over an X-band patch antenna and a 100 times increase over state-of-the-art S-band parabolic antennas. This paper discusses designing, building, integrating, and operating the flight antenna from a mechanical perspective, its final performance, and lessons learned. KaPDA enabled the RainCube mission, the first Earth Science CubeSat to have an active instrument. RainCube was launched in May 2018, making KaPDA the second deployable parabolic antenna to fly on a CubeSat and the first to operate in Ka-band, enabling follow-on opportunities for high-rate antenna communications and remote sensing science.

Recent studies on neural networks in solving Maxwell's equations inquire about the explainability of neural networks. Most neural networks are constructed by multiple neural layers, which are in a similar architecture as numerical iterative solvers for Maxwell's equations. This paper demonstrates the work that launches a fast iterative solver for integral equations as a neural network on a neural network platform, PyTorch. The fast iterative solver is the Conjugate-Gradient Fast-Fourier-Transform (CGFFT) solver, considering its fast convergence, small computation cost, and good scalability. A 1D single-slab scattering problem and a 2D cylinder scattering problem are solved through the CGFFT neural network implemented on PyTorch. With PyTorch's optimized computation scheme for matrix computation and GPU parallelization, the solver takes much less time compared to its traditional implementation.

Particle swarm optimization (PSO) is a well-known optimization technique in electro magnetics. To ensure particles remain within the designated solution space throughout the optimization procedure, various boundary conditions are employed in PSO algorithms. However, there are instances when the global optimum is not present within the specified solution space. Consequently, regardless of how extensively the particles search, they cannot achieve the desired fitness value. This paper introduces an adaptive solution space method that efficiently expands the solution space to attain the desired fitness goals. The performance of the adaptive solution space method is tested using both mathematical benchmark functions and a real-world electromagnetic problem. Simulation results confirm the efficiency and effectiveness of the proposed adaptive solution space method.

The switch-reconfigured unit cell (SRUC) capable of generating variable phase response is one of the enabling technologies for reconfigurable intelligent surfaces (RIS) and beam steerable reflect-/transmit-arrays. Traditional approaches designing SRUC have primarily relied on basic element geometries with specific switch placement. However, these methods lack design freedom and result in excessive use of switches, limitations in performance, and complexity in design. In this paper, we employ binary particle swarm optimization (BPSO) to simultaneously design the element topology and switch placement of the SRUC. Specifically, the design and performance of a 4-state (2-bit) K-band circularly polarized (CP) reflectarray SRUC that utilizes single-layer element and merely two switches is presented.

The cosine-q radiation pattern formulas allow one to create sidelobe-free patterns with easily controlled beamwidth and taper. These advantages make the cosine-q pattern an ideal feed source in the studies of reflector, transmit-/reflect-array, and lens antennas. However, in this paper, it is shown that the original cosine-q formulas are limited to rotationally symmetric patterns and lose their effectiveness when an elliptical pattern (or fan-beam) with highly unequal beamwidths in two principle cuts are desired. This limits the application of cosine-q sources in antennas with elongated apertures, which are of great interest in radar sensing and CubeSat applications. This paper proposes a set of new formulas that enable the use of cosine-q-type illumination for elongated aperture antennas. The model is further improved by adding a phase function that accounts for the interleaved phase center locations, which can appear in practical feed sources. The effectiveness of the new cosine-q formulas is demonstrated through analyzing two representative antennas: a C-Band linearly-polarized 7 m×1.5 m aperture reflector antenna and a K-band circularly-polarized 20 cm×10 cm reflectarray antenna.

In this article, we present a systematic analysis on the frequency selection scheme of wireless link, toward miniaturized backscattering neural implants. A novel frequency selection scheme is proposed that takes more comprehensive consideration of frequency-dependent design constraints, including the power transfer efficiency (PTE) of the coils, the limitations of human electromagnetic exposure, the RF-to-dc conversion efficiency, and the power budget of a backscattering circuit. Frequency-dependent design constraints of brain implants are critically investigated first. Two representative application scenarios are then analyzed accordingly, including implants at the upper limb peripheral nervous system (PNS) and spinal cord. It is concluded that an upper limit of the wireless link operating frequency exists for miniaturized backscattering implants. For the application of cortical interfaces, the feasibility of the scheme was verified through prototyping and measurements at representative frequencies.

The IEEE Society’s 140th anniversary and the IEEE Antennas and Propagation Society’s 75th anniversary mark significant milestones in their enduring academic and industrial contributions. Notably, there is a discernible rise in demand for high-performance antennas, continually challenging the boundaries of manufacturing techniques. Innovations in manufacturing act as catalysts, propelling researchers and engineers to explore unconventional design paradigms and push modern antenna performance limits. The evolution of printed lens and transmitarray antenna technologies exemplifies collaborative out-of-the-box developments at the intersection of antenna design and manufacturing methods. In this invited review article, we delineate key advancements in lens and transmitarray antenna manufacturing technologies and their consequential impact on antenna design methodologies. Commencing with foundational techniques, such as milling and injection molding, advancing through printed circuit board (PCB) processes, and culminating in the cutting-edge capabilities of 3D printing, this article reviews the chronological progression that significantly enhances the design possibilities and functionality of lens and transmitarray antennas. Ample representative references are provided to facilitate further exploration of research conducted by other researchers.