Giuseppe Thadeu Freitas de Abreu is a leading figure in modern wireless communications, whose work spans communications theory, signal processing, and advanced electromagnetic systems. Currently serving as Scholar in Residence at Digital Futures (September 2025 – January 2026), he brings decades of international academic experience across Brazil, Japan, Finland, and Germany, alongside deep engagement with both theoretical foundations and emerging technologies.
With research interests ranging from integrated sensing, communications, and computing to reconfigurable electromagnetic structures such as intelligent metasurfaces and XL/Gigantic MIMO, Giuseppe has played a key role in shaping how next-generation wireless systems are conceptualized and optimized. His contributions extend beyond research, having served in several editorial leadership roles within the IEEE and receiving multiple international awards for his work.
At a Digital Futures seminar on 20 November 2025, Giuseppe Abreu, spoke about “A Discrete Twist on the Linear Inversion Problem”. Here’s a link to the recorded presentation on Youtube.
In this interview, we explore his academic journey, current research directions at Digital Futures, and his perspective on the future of wireless communications—both from a technological and a research culture standpoint.
Your academic journey spans Brazil, Japan, Finland, and Germany, and bridges engineering, physics, and signal processing. How have these interdisciplinary and international experiences shaped the way you approach research in modern wireless communications today?
– I think international experience affects educational and research mentoring aspects more than research itself. That said, I have indeed been fortunate throughout my academic path to have been exposed to different academic systems, which strongly influence how I approach research mentoring in wireless communications—especially when it comes to introducing students to a research career in the area.
To make a comparison, in Brazil, the B.Sc. degree in Electrical Engineering alone is a five-year program, with the first two years largely overlapping with mathematics and physics curricula, thus providing a rigorous and strongly foundational engineering education. However, research is not systematically integrated into undergraduate studies, which often delays students’ exposure to the possibility of a career in science. This experience taught me the importance of involving talented students in research early—sometimes as early as the second year—once they have acquired core tools such as linear algebra, probability, optimization, and programming. In my research group, I therefore actively engage young students very early through well-scoped mini research projects, an approach that has consistently paid off. I have had several brilliant students publish high-quality articles during their undergraduate studies and skip the master’s degree to go straight into a PhD.
In Japan, by contrast, undergraduate programs are four years long, with research formally integrated into the final year. However, perhaps because universities are so large, with extensive faculty bodies offering a wide range of topics, the first three years often lack structure. Students therefore spend those foundational years exploring before choosing a thesis topic, often without sufficient depth in the relevant background. As a result, much of the final year is spent learning fundamentals rather than producing results. This partially explains why, in Japan, the vast majority of students stay at the same university to pursue a master’s degree, as most want to finish what they started during the bachelor’s program. While this system promotes early research exposure, paradoxically it also contributes to longer study paths, at high cost to students (six years of tuition fees, even in public institutions), who are then pushed toward industry for financial rather than vocational reasons. From this, I learned the value of structured curricula with clearer academic tracks, which help students find focus earlier without discouraging exploration.
In Europe, my opinion is that the Bologna framework provides a balanced structure, with strong industry–academia collaboration, particularly in wireless communications in countries such as Finland, Germany, and of course Sweden. All of this is excellent, but one area where Europe could improve somewhat is integration toward doctoral studies. At Constructor University, for instance, we address this with a fast-track PhD option for outstanding students, allowing earlier entry into doctoral research. I have had several students graduate with a PhD as young as 24 years old. This gives them the freedom to pursue ambitious paths; several of my former PhD students have gone on to found successful start-up companies, both in Finland and Germany. I wish Europe adopted this approach more systematically, as in the U.S. Science must lead to innovation, especially in engineering.
Finally, regarding interdisciplinarity, I find it most impactful at earlier career stages, as it helps identify interests and strengths, while depth naturally becomes more important as researchers mature. At the team level, however, interdisciplinary diversity is crucial and is one of the aspects that makes engineering research particularly dynamic and rewarding.
At Digital Futures, your work focuses on Integrated Sensing, Communications, and Computing (ISCC). What do you see as the main conceptual or practical barriers to truly integrating these functions into a unified wireless system, and how close are we to real-world deployment?
– I am convinced that future wireless systems must go beyond communications and incorporate other functionalities such as sensing, computing, and eventually control. This is due to a fundamental shift currently underway: future wireless systems will primarily serve AI-empowered machines rather than humans. In 2025, Internet traffic due to bots surpassed that of humans, at 51% versus 49%. This trend will continue because, unlike humans—who rely on different senses such as vision, hearing, smell, and touch, and therefore different signals—to interact with the world, machines can do so using the same electromagnetic waves. That is what ISCC is all about.
So the market is clearly there. There is also a precedent dating back to the launch of 4G systems, which integrated localization capabilities, giving rise to location-based services (LBS). Localization can be seen as an early form of ISCC since, similar to sensing, it is performed using “meta-information,” i.e., not data itself, but variations in received signal strength, phases, and Doppler shifts introduced during propagation and detected at the receiver. In turn, the location of a device is an aggregate function of this meta-information, which is conceptually similar to over-the-air computing (OTAC). ISCC is a generalization of this idea, where sensing extends to multiple environmental parameters—e.g., orientation, size, and speed—and computing includes the ability to evaluate a wide range of nomographic functions based on aggregation.
The global LBS market is expected to grow to 236 billion dollars by 2033. Imagine how much larger the broader ISCC market will be. Industry interest is therefore strong, but predicting time to market is difficult, as it depends not only on technological maturity and market potential, but also on factors such as standardization, IP consolidation, economic conditions, and investment capacity. That said, there are already encouraging signs: standardization of integrated sensing and communications (ISAC) is underway, both in Wi-Fi, through IEEE 802.11bf, and in cellular systems as part of ongoing discussions toward 6G. The European Telecommunications Standards Institute (ETSI) has also recently created a work item on integrating OTAC with ISAC, which effectively amounts to ISCC.
Finally, I would not describe the remaining challenges as barriers, since I see no fundamental deal-breakers. Rather, there are challenges arising from the distinct requirements imposed by multifunctionality on wireless systems. For example, communication requires unknown signals to be separated at the receiver, while computing requires their aggregation. Although in both cases signals that are impervious to the propagation environment are preferred to minimize distortion, sensing demands signals that are sensitive to environmental effects to maximize accuracy. Moreover, while sensing benefits from prior knowledge of transmit signals at the receiver, this is incompatible with the communication objective. These conflicting requirements translate into research questions on performance trade-offs, waveform design, and receiver architectures capable of supporting all functionalities simultaneously. These are precisely the challenges currently being addressed by the research community, including ourselves.
Metasurfaces and XL/Gigantic MIMO are a central part of your current research. In your view, how transformative are these technologies for future wireless networks, and what trade-offs are still underestimated by the research community?
– Indeed, these technologies—together with continuous aperture arrays (CAPA), dynamic scattering arrays (DSA), and reconfigurable intelligent surfaces (RIS), sometimes grouped under the term reconfigurable electromagnetic structures (RES)—represent key innovations for future wireless systems. Setting aside the specific features of each approach, they share the ability to configure parameters to optimize their response to electromagnetic waves and enhance system key performance indicators (KPIs). The mathematical problems associated with such optimization are among my main research interests.
Focusing on metasurfaces specifically, these are artificially engineered two-dimensional materials whose properties can be optimized at the nano- or atomic scale to achieve unprecedented control over electromagnetic waves. In other words, the number, type, and effect of their parameters can be enormous. This enables the design of highly sophisticated RES with advanced features, lower cost, and high energy efficiency. A direct consequence is the possibility of performing “wave-domain signal processing,” where operations such as beamforming, modulation, filtering, and direction-of-arrival estimation are carried out directly through wave–material interaction, reducing hardware complexity and processing latency, potentially at the speed of light. I therefore strongly believe this technology has transformative potential, an opinion widely shared in the community and reflected in the intense recent research activity on the topic.
Regarding trade-offs and open research problems, there is indeed much work to be done. One key issue is response time. While metasurfaces can, in principle, perform speed-of-light wave-domain processing, switching parameters from one state to another requires time. For example, the typical switching speed of liquid crystal displays (LCDs)—a current technology used to build metasurfaces—is on the order of tens of milliseconds, which is too slow for high-data-rate wireless systems. Recent research has demonstrated switching times in the microsecond and even nanosecond range, so progress is clearly being made. Another challenge lies in the computational complexity of optimizing large metasurfaces, which can be prohibitive, although quantum computing may help mitigate this issue. Finally, accurate physics-compliant models are still lacking, as much of the current literature relies on simplified assumptions that do not fully match experimental observations. The research community is actively addressing these challenges, and we are contributing as well.
You have contributed extensively as an editor and reviewer for leading IEEE journals. From this perspective, how have research priorities in wireless communications evolved over the past decade, and what emerging themes do you believe will define the next one?
– To highlight the evolution, let me compare two periods: 2005–2015 and 2015–2025. For conciseness, I will focus on cellular systems, leaving aside other important developments such as the rise, fall, and resurgence of ultrawideband (UWB) technology, advances in wireless positioning, and the Internet of Things (IoT).
In the first period, the community focused on refining 4G and laying the foundations for 5G. Two dominant topics were orthogonal frequency-division multiplexing (OFDM), which replaced the code-division multiple access (CDMA) technology of 3G, and multiple-input multiple-output (MIMO) systems, along with related topics such as beamforming and space–time coding and multiplexing. Together, these technologies aimed to increase data rates and user capacity by better managing interference.
While these objectives remain relevant today through concepts such as cell-free architectures and massive MIMO—areas in which my Digital Futures host, Prof. Emil Björnson, is a leading contributor—new goals and technologies have emerged as equally, if not more, important for 6G and beyond. As discussed earlier, multifunctionality (e.g., ISAC and ISCC) is one such goal, and metasurfaces represent a key enabling technology. Beyond these, two other themes of major importance are artificial intelligence (AI) and quantum technologies.
Regarding AI, I believe future wireless systems must not only use AI for design and operation but be fundamentally conceived for intelligent machines. This mirrors the transition from voice-centric systems in 1G and 2G to today’s data-dominated networks, where video traffic accounts for roughly 80% of mobile data. Similarly, future networks will primarily serve machines and support functionalities beyond communication.
As for quantum technologies, their impact may arrive sooner than many expect. Although quantum computing is often viewed as applicable only to niche problems, history shows that new computational paradigms tend to find broader applications once they become accessible. When practical and affordable quantum technologies emerge, their influence on wireless systems and signal processing will be profound.
For early-career researchers entering the field today, what skills, mindsets, or research directions would you recommend to remain relevant in such a fast-moving and increasingly interdisciplinary landscape?
– Many of my recommendations can be inferred from the previous answers, but I would place strong mathematical foundations, critical thinking, and AI literacy at the top of the list. The latter may be controversial, as some worry that AI diminishes research quality, but I disagree. The role of a scientist is to solve problems and innovate, and this involves many time-consuming tasks. AI helps reduce the time spent on some of these tasks.
For example, we used to spend days in libraries searching for relevant articles. Today, this can be done in seconds online, and it would be absurd to argue that literature reviews are now of lower quality as a result. On the contrary, the time saved can be invested in deeper reading and thinking. Similarly, AI can accelerate programming, verification of derivations, and aspects of technical writing, freeing time for creativity and insight. Like any powerful tool, it must be used wisely, but when used correctly, it benefits rather than harms science.
