IEEE IMAS 2026 Computational Electromagnetics Workshop

Francesco P. Andriulli received the Laurea in electrical engineering from the Politecnico di Torino, Italy, in 2004, the MSc in electrical engineering and computer science from the University of Illinois at Chicago in 2004, and the PhD in electrical engineering from the University of Michigan at Ann Arbor in 2008.

From 2008 to 2010 he was a Research Associate with the Politecnico di Torino. From 2010 to 2017 he was an Associate Professor (2010-2014) and then Full Professor with the École Nationale Supérieure Mines-Télécom Atlantique (IMT Atlantique), Brest, France. Since 2017 he has been a Full Professor with the Politecnico di Torino, Turin, Italy.

His research interests are in computational electromagnetics including frequency- and time-domain integral equation solvers, well-conditioned formulations, fast solvers, low-frequency electromagnetic analyses, and modeling techniques for antennas, wireless components, microwave circuits, and biomedical applications with a special focus on brain imaging.

Prof. Andriulli received several best paper awards at conferences and symposia and numerous other awards and recognitions for his research contributions.

He received the 2014 IEEE AP-S Donald G. Dudley Jr. Undergraduate Teaching Award, the triennium 2014-2016 URSI Isaac Koga Gold Medal, and the 2015 L. B. Felsen Award for Excellence in Electrodynamics.

Prof. Andriulli is a Fellow of the IEEE and of the International Union of Radio Science (URSI). He serves as the 2026 President-Elect of the IEEE Antennas and Propagation Society.

Computational Electromagnetics (CEM) is a scientific discipline positioned at the intersection of electrical engineering, high-performance computing, and applied mathematics. It is primarily concerned with the modeling and simulation of complex electromagnetic phenomena that arise in advanced technological and scientific contexts.

Historically, CEM has provided the predictive framework underlying a wide spectrum of applications in electrical and electronic engineering, optics, wireless communications, geophysical exploration, and biomedical systems.

In recent years, the scope of CEM has broadened significantly, engaging with an expanding range of interdisciplinary fields, including information retrieval, computational neuroscience, machine learning, and brain-computer interfaces.

A pervasive trend across these domains is the substantial growth in the dimensionality and scale of the problems to be addressed, driven by system miniaturization and increased component density. This evolution has led to escalating levels of computational complexity and, consequently, increased cost and effort associated with electromagnetic modeling, design, and simulation.

This talk will provide a comprehensive overview of the major challenges, both longstanding and emerging, facing the field of CEM, with examples spanning from canonical problems to those of “Holy Grail”-like complexity.

Emphasis will be placed on recent research efforts aimed at achieving paradigm shifts in simulation methodologies, with the objective of overcoming the computational bottlenecks inherent in conventional approaches and enabling significant performance gains.

Applications discussed will include brain imaging and modeling, electromagnetic dosimetry, epilepsy diagnosis, and the development of brain-computer interfaces.

Simon B. Adrian (S’09–M’19–SM’23) received the Bachelor of Science (B.Sc.) degree in Electrical Engineering and Information Technology from the Technische Universität München (TUM), Munich, Germany, in 2009.

He received the Master of Science (M.S.) degree in Electrical and Computer Engineering from the Georgia Institute of Technology, Atlanta, GA, USA, in 2010, the Diplom-Ingenieur (Dipl.-Ing.) degree in Electrical Engineering and Information Technology from the TUM in 2012, and the Doktor-Ingenieur (Dr.-Ing) degree from the TUM and the École nationale supérieure Mines-Télécom Atlantique Bretagne Pays de la Loire (IMT Atlantique), Brest, France, in 2018.

From 2012 to 2019, he was a Research Assistant at the TUM. In 2018, he was Visiting Professor with the Politecnico di Torino, Turin, Italy.

From 2019 to 2020, he was a Senior Engineer with Infineon AG, Neubiberg, Germany. From 2020 to 2024, he was an Assistant Professor with the Universität Rostock, and since 2024 he has been a Professor.

His research interest is in computational electromagnetics, focusing on integral equation solvers. In particular, he is interested in preconditioning techniques, low-frequency stable formulations, and fast solvers. Areas of application include antenna modeling and bio-electromagnetic problems.

Dr. Adrian has been a member of the IEEE Antennas and Propagation Society Education Committee since 2016, where he has been chair of the Student Activities Subcommittee since 2021.

He has been an Associate Editor of the IEEE Transactions on Antennas and Propagation since 2020 and the IEEE Antennas and Propagation Magazine since 2021.

He has received several research and paper awards, including the IEEE APWC-ICEAA Award 2022, the Young Scientist Best Paper Award of the Kleinheubacher Tagung 2024, and the 2nd prize of the Karl-Jörg Langenberg Award at the Kleinheubacher Tagung 2025.

EEG from scalp potentials, and its source-reconstruction counterpart, is a mature and widely used modality for activity tracking and functional brain imaging.

However, high-resolution source reconstruction remains fundamentally limited by the accuracy and scalability of the underlying electromagnetic models, as well as by persistent computational bottlenecks.

In focal epilepsy, these limitations directly affect source localization in presurgical workflows; in BCI applications, they constrain spatial resolution, robustness, and real-time performance.

High-resolution EEG systems are, in fact, computationally demanding, as a substantial part of the imaging pipeline relies on advanced electromagnetic models of signal propagation in the head.

For this reason, innovations in computational methods, modeling strategies, and algorithmic approaches have become central to cross-disciplinary research at the intersection of Computational Electromagnetics (CEM), neuroengineering, and applied electromagnetics.

This talk will present recent advances in CEM-enabled neuroimaging, focusing on technologies for brain diagnosis, therapy, and interaction, where improved computational methods and dedicated platforms enable more accurate and scalable solutions.

Current trends and open Grand Challenges will be discussed alongside established results and ongoing research activities, including those carried out within the ERC project TurboEEG and the EIC Pathfinder project CEREBRO.

The talk will highlight recent theoretical and experimental developments and relate them to applications in diagnostics, immersive neurofeedback and brain–machine interfaces.

A variety of fast methods has been developed to accelerate the solution of surface integral equations discretized using the boundary element method.

For electrically small problems, where the wavelength is larger than the underlying geometry, algebraic—and therefore kernel-independent—low-rank factorization methods, such as the adaptive cross approximation (ACA) for the construction of H-matrices or the nested cross approximation (NCA) for the construction of H²-matrices, can be utilized.

These methods exploit the fact that, while the system matrix is generally full rank, blocks of the system matrix associated with well-separated parts of the geometry are approximately low rank.

As the frequency increases and the problem becomes electrically large, the rank of these blocks can grow as the mesh is simultaneously refined, rendering these techniques less attractive.

To address this limitation, directional compression techniques, such as directional H²-matrices, have been introduced to exploit the directional low-rank structure of high-frequency interactions.

Directional H²-matrices, originally introduced for Helmholtz-type operators, allow a compressed representation of the system matrix to be constructed using low-rank approximation techniques with log-linear complexity.

The construction of H²-matrices begins with a hierarchical partitioning of the underlying geometry, followed by a block decomposition of the system matrix into compressible and noncompressible blocks.

For the construction of directional H²-matrices, an additional directional hierarchy is introduced into the geometrical partitioning.

The computation of the nested low-rank factorization of the compressible blocks is typically performed using ACA. While ACA often achieves a near-optimal factorization, only a small portion of the resulting factorization is actually used in the H²-matrix.

To overcome this limitation, the incomplete adaptive cross approximation (IACA) reduces the unused portion of the factorization and thereby significantly improves overall performance.

Recent progress on the IACA for standard and directional H²-matrices applied to the electric field integral equation (EFIE) demonstrates the benefits of using an incomplete ACA algorithm compared with the standard ACA algorithm.

Abdulkadir C. Yucel is an Assistant Professor in the School of Electrical and Electronic Engineering at Nanyang Technological University (NTU), Singapore.

His research lies at the intersection of computational electromagnetics, electromagnetic sensing and imaging, and artificial intelligence, with a particular focus on physics-informed AI for intelligent radar systems and fast electromagnetic simulation.

Dr. Yucel received his Ph.D. in Electrical Engineering from the University of Michigan, Ann Arbor, and conducted postdoctoral research at the University of Michigan, the Massachusetts Institute of Technology (MIT), and King Abdullah University of Science and Technology (KAUST).

His recent research includes AI-powered radar systems for non-invasive sensing and imaging, deep learning- based electromagnetic surrogate models, and fast computational electromagnetics methods based on FFT, FMM, butterfly, and tensor techniques.

He is a recipient of the Fulbright Fellowship (2006), the IEEE Transactions on Power Electronics Prize Paper Award (2024), and the ACES Technical Achievement Award (2026).

He is a Senior Member of IEEE and serves as an Associate Editor for IEEE Antennas and Propagation Magazine and the IEEE Journal on Multiscale and Multiphysics Computational Techniques.

He has authored over 160 peer-reviewed publications.

Electromagnetic technologies underpin a wide range of sensing, imaging, biomedical, and integrated-circuit applications.

However, their practical deployment is often limited by two fundamental challenges: reliable interpretation of electromagnetic measurements in complex environments and the high computational cost and memory requirements of accurate electromagnetic simulation.

In this talk, I will present recent advances in physics-informed artificial intelligence that address these challenges through the integration of electromagnetic physics, signal processing, computational methods, and learning.

The first part of the talk focuses on intelligent radar sensing and imaging, with emphasis on radar systems for non-invasive inspection.

I will present our recent work on AI-powered tree-trunk radar for rapid defect detection and imaging and tree-root radar for three-dimensional subsurface mapping.

These systems combine custom-designed radar hardware and automated data acquisition with physics-guided signal processing and learning-based detection and inversion to achieve rapid and robust reconstruction under practical field conditions.

The second part focuses on fast computational electromagnetics. I will present deep learning-based electromagnetic surrogate models that enable near-real-time prediction of electromagnetic responses for applications including ground-penetrating radar, biomedical electromagnetics, and integrated circuits.

I will also discuss fast physics-based computational electromagnetics solvers accelerated through FFT, FMM, butterfly, and tensor techniques for large-scale electromagnetic analysis and efficient generation of high-fidelity training data for AI models.

Together, these developments illustrate how physics-informed AI and fast computational electromagnetics can work synergistically toward intelligent, accurate, and real-time electromagnetic sensing and simulation.