Applied Computational Electromagnetics Society Journal (ACES)

A Comparative Study of Anti-Jamming Beamforming Using Deep Learning in Planar Phased Array Antennas

Thu, 30 Apr 2026 00:00:00 +0200

In this study, a deep learning-based beamforming comparative study for anti-jamming applications in 2D-planar phased arrays is presented. For better array architecture benchmarking, three different geometries (circular, rectangular, hexagonal) are considered. Convolutional Neural Network (CNN) is employed to translate a target radiation pattern, generated as an image, directly into the optimal antenna currents. Adaptive antenna array beamforming weights can be estimated efficiently by the deep learning-based MATLAB code according to the desired beam steering angle and the null direction of the jammer. This approach establishes a smart, non-iterative mapping that bypasses traditional optimization algorithms, reducing computation time by up to 260x. Once trained, the model delivers optimal currents and weights in a single and efficient forward pass.

Air-to-Ground Path Loss Modeling in UAV Networks Via GSA-Based Hyperparameter Optimization

Thu, 30 Apr 2026 00:00:00 +0200

In Unmanned Aerial Vehicle (UAV) communications, Air-to-Ground (A2G) channel modeling is complex due to high mobility and environmental dynamics. While Machine Learning (ML) and Deep Learning (DL) techniques have been adopted to improve prediction accuracy over traditional empirical models, their performance remains highly dependent on hyperparameter configuration. Recent techniques such as Random Search and Bayesian Search are commonly used for hyperparameter tuning; however, they often struggle with convergence efficiency and prediction stability. To address these challenges, this study aims to develop a hyperparameter tuning framework based on the Gravitational Search Algorithm (GSA) to enhance the predictive performance of ML-based A2G models. The framework is applied to K-Nearest Neighbors (KNN), Decision Tree (DT), Random Forest (RF), and Long Short-Term Memory (LSTM) models at 1 GHz, 2 GHz, and 5.8 GHz. Experimental results demonstrate that GSA-optimized models demonstrate improved predictive stability and competitive accuracy, with GSA-LSTM and GSA-RF achieving an Root Mean Square Error (RMSE) of 5.46 dB, representing a $56 %$ improvement over the free-space model. The proposed approach demonstrates improved robustness compared to conventional search strategies.

Optimization of a Wideband Rectangular TEM Device by Genetic Algorithms

Shiqi Wang, Yangyi Fu, Jinyu Deng, Guojie Wang, Jiayu Sun — Thu, 30 Apr 2026 00:00:00 +0200

In recent years, artificial intelligence has been widely introduced into the design of electromagnetic devices. Traditional designs of DC-5.2 GHz wideband rectangular transverse electromagnetic (TEM) devices depend on complex formulas and electromagnetic simulation software such as HFSS and CST Microwave Studio Suite TM 2013. This paper proposes a DC-5.2 GHz rectangular TEM device optimized by genetic algorithms (GAs). The main innovation is the comparison between AI-based optimization and traditional design methods while ensuring excellent wideband transmission performance. The GA-optimized TEM device presents favorable performance and is suitable for cellular radiation experiments in wireless communication systems.

Design of All-Dielectric Resonant Metasurface Receiver for Millimeter-Wave Front-Ends

Thu, 30 Apr 2026 00:00:00 +0200

In recent years, high-power millimeter waves (HPMMW) have emerged as a severe electromagnetic disturbance that endangers telecommunication networks significantly, as the intense electromagnetic interference induced by HPMMW can damage metallic electronic circuits and front-end devices. To address this issue, this paper proposes an all-dielectric resonant metasurface-based front-end system that enables millimeter-wave components to resist HPMMW-induced damage. The system comprises a laser, an all-dielectric metasurface, an electro-optic resonator, a photodetector, and electronic circuitry. A $3 \times 3$ cell-array metasurface antenna is used to capture and transmit a 65 GHz millimeter-wave signal to an electro-optic field sensor. Subsequently, the photodetector converts the optical signal into a demodulated radio-frequency signal, which can be further processed by the subsequent electronic circuitry. With a compact footprint of $7.7 \times 7.7 {mm}^{2}$ and a high receiver sensitivity of $-$ 52 dBm, the proposed system can be integrated with other electronic circuits, facilitating the miniaturization of telecommunication equipment.

Design of an Electronically Continuous Tunable Phased Array Antenna

Chang-Keng Lin, Ding-Bing Lin, Chang-Ching Lin — Thu, 30 Apr 2026 00:00:00 +0200

In this paper, an electronically continuous tunable phased array antenna is proposed, which integrates a 1-port to 4-ports unequal power divider, four electronically continuous tunable phase shifters, and a $4 \times 3$ right hand circularly polarized patch array antenna. The unequal power divider is designed with a power ratio of 1:2.25:2.25:1 to achieve a side-lobe level suppression of 20 dB. The phase shifters provide electronically linear and continuously tunable phase control within $\pm 180^{\circ}$ . Their phase difference can be easily controlled by adjusting the direct current voltage. The array antenna consists of 12 hexagon patches, providing right hand circular polarization within the operating bandwidth and achieving a high gain of 13.68 dB. Furthermore, the proposed phased array antenna is capable of continuously steering the main beam over a range of $- 50^{\circ}$ to $55^{\circ}$ in the Y-Z plane at 3 GHz. Finally, the measurement results show good agreement with the simulations, confirming that the proposed electronically continuous tunable phased array antenna exhibits excellent performance.

A Substrate-Loaded Gain-Enhanced Vivaldi Antenna Design for the SolidState High-Power Microwave Module

Zichong Chen, Fangsheng Cai, Peng Bai, Taijing Shi, Xiaojun Mao, Yun Jiang — Thu, 30 Apr 2026 00:00:00 +0200

This paper presents an evolutionary design process, fabrication, and verification of a substrateloaded Vivaldi antenna (VA) design for the pulse-type high-power microwave (HPM) solid-state T/R module. The antenna design utilizes a substrate with high permittivity to improve its power handling capacity (PHC), a snowflake-like metasurface (SFL-MS) lens on the front of the VA to enhance the directivity and impedance characteristics in the middle and high frequency bands, and rectangular slits on the radiating brims of the VA for further broadening its bandwidth and realized gain at low frequencies. After HPM measurement, these functional methods are proven to be effective for jointly contributing to optimizing antenna performances. The proposed prototype shows an operating band of 2.28–6.54 GHz (voltage standing wave ratio [VSWR] $<$ 2) and the PHC values are over 8000 W in this band. The measured realized gain and the maximum gain enhancement can reach 6.56–8.15 dBi and 5.96 dBi. The measured results are reasonable and agree well with simulations.

A Miniaturized Four-Port MIMO Slotted Microstrip Patch Array Antenna Design With Reduced Mutual Coupling for 5G Wireless Applications

Venkatrao Kolli, Merlin Sheeba G — Thu, 30 Apr 2026 00:00:00 +0200

In this paper, a four-port Multiple-Input Multiple-Output (MIMO) antenna array is designed and analyzed for millimeter-wave (mm-wave) 5G applications. The configuration comprises a slotted microstrip patch array antenna excited through a T-junction power divider/combiner-based feed network. The set of truncated square slots are arranged on a rectangular patch to act as radiating surfaces. Similarly, the ground plane is a combination of rectangular and square-shaped slots used in a ground plane to enhance impedance bandwidth and radiation characteristics. The proposed fourport antenna is fabricated and the prototype is experimentally characterized for the S -parameters ( $S_{11}, S_{21}, S_{31}, S_{41}, S_{32}, S_{42}, S_{43}$ ), radiation pattern, and gain. Typical dimensions of the antenna are $28 \times 28 \times 0.254 {mm}^{3}$ . Measured results are in excellent agreement with simulated results. The four-port MIMO antenna operates effectively over the 27.45–28.55 GHz frequency range and is suitable for emerging 5G applications with a gain of 10.02 dBi. To introduce the isolation between the adjacent elements in the array, the technique of polarization diversity has been employed. The corresponding enveloper correlation coefficient (ECC) suppression has been noticed. Based on the results, the proposed antenna confirms excellent diversity performance, and hence the design can be a promising solution for mm -wave and 5G applications.

Gain-Focusing Performance Evaluation of Nonlinear Frequency Diverse Arrays

Xin Wang, Tao Jiang, Chengkai He, Gengzuo Liu — Thu, 30 Apr 2026 00:00:00 +0200

Nonlinear frequency diverse arrays have attracted increasing attention because of their unique point-like focusing characteristics in the joint angle-range domain. However, the quantitative relationship between the focusing pattern and key design parameters, as well as systematic performance evaluation tools, have not been fully established. In this paper, we propose the gain-focusing area as a quantitative metric for assessing the focusing efficiency of nonlinear frequency diverse arrays (NFDA). Unlike conventional one-dimensional beamwidth measures, the gain-focusing area captures the joint angle-range behavior of the array and provides a basis for performance evaluation and design optimization. Using this metric, we systematically study single- and dual-region focusing, and quantify the influence of focusing location (angle, range) and array design parameters (number of elements, frequency-offset coefficient) on the gain-focusing performance. Numerical simulations demonstrate that the proposed gain-focusing-area based evaluation method provides a more reliable performance metric; compared with existing approaches, it nearly doubles the valid angular range while maintaining accurate characterization of the focusing behavior. This enables NFDA spatial focusing capability and parameter sensitivity to be robustly quantified over a wider field of view, offering a more dependable tool for shaping the electromagnetic environment in and around the target region.

Synthesis and Design of a Wideband Filtering Impedance Transformer and its Application as a Power Divider

Haili Zhang, Zhenzhong Chen, Taijun Liu — Thu, 30 Apr 2026 00:00:00 +0200

In this paper, two classes of impedance transformers with wide operating bandwidth are presented. The impedance parameters can be directly determined based on the specified impedance transformation ratio, fractional bandwidth (FBW), and in-band return loss ( $R L$ ) requirements. To validate the proposed method, an impedance transformer is designed at the center frequency $(f_{0})$ of 2.4 GHz and an impedance-transforming ratio $(r)$ of 0.5. The filtering impedance transformer with FBW $=$ 90% and $R L = 20 dB$ is fabricated and measured, validating the theoretical prediction. Furthermore, to demonstrate its application potential, the second impedance transformer is employed to realize a filtering power divider with wide operating bandwidth and enhanced isolation bandwidth.

Development of Mesh-Based Generated Reluctance Network Using Trapezoidal Elements Based on Lumped Parameter Model

Thu, 30 Apr 2026 00:00:00 +0200

This paper develops a novel Mesh-Based Generated Reluctance NetWork (MBGRN) model, which is based on the lumped parameter modeling method. The mesh-based approach automates network generation, replacing the manual flux path definitions required in traditional magnetic equivalent circuit (MEC). In this approach, the computational domain is represented in a polar coordinate system where the mesh elements are defined as isosceles trapezoids. The model utilizes a rotation simulation to bypass the remeshing processes common in the finite element method (FEM). A key advantage of the proposed MBGRN method is that the number of computational elements is reduced by half compared to the conventional FEM. This leads to a significant reduction in computation time, ranging from 10 to 15 times faster than traditional FEM, while maintaining a calculation error of less than 1% relative to the FEM. The development of this method is validated through a practical benchmark problem: the surface-mounted permanent magnet synchronous motor under no-load condition. The results obtained from the MBGRN model are thoroughly compared with those from the 2D FEM.