The study addresses a structural constraint in conventional radio access networks, where uplink and downlink operations are tightly coupled and scheduling relies heavily on explicit channel-state information feedback. Decoupling those directions gives base stations greater freedom in coordinating resource allocation because the scheduling process does not require the same feedback-based information exchange. Liu and colleagues formulated resource-scheduling problems for this architecture under two network regimes, with one algorithm designed for heavy-load conditions and another for light-load conditions.

For heavy-load states, the team developed an optimal greedy algorithm that maximizes the weighted sum of user demand-satisfaction rates. The algorithm therefore provides a formal optimization objective for allocating scarce resources under heavy demand. The authors report that simulations showed superior performance and running time for their proposed algorithms, although the supplied article abstract does not specify the numerical margins or the range of scenarios tested.

For light-load states, the researchers proposed an optimal two-stage scheduling algorithm. This approach is designed to meet all user demands while minimizing network energy consumption by using the flexibility available in the decoupled architecture. The distinction between the two algorithms reflects their different objectives: the heavy-load method maximizes the weighted sum of demand-satisfaction rates, whereas the light-load method seeks full demand satisfaction with lower energy use.

Together, the formulations show how independently operating uplinks and downlinks can change the scheduling choices available to cooperating base stations. The prototyping implementation provided an additional check on the proposed feedback-free approach, with the authors reporting that it validated the method’s practical feasibility. The evidence is nevertheless bounded: the reported results combine network simulations with a prototype rather than field measurements from an operating sixth-generation network.

Its central technical result is that decoupling uplink and downlink operations can make feedback-free multi-base-station resource scheduling possible within the modeled framework.

Subject of Research: Feedback-free multi-base-station resource scheduling in fully decoupled 6G radio access networks

Article Title: Flexible multi-base-station cooperation towards 6G: a fully decoupled feedback-free access approach

Article References: Liu, J., Chen, J., Shi, Y., Sun, Z., Qian, B., & Zhou, H. (2026). Flexible multi-base-station cooperation towards 6G: a fully decoupled feedback-free access approach. Communications Engineering. https://doi.org/10.1038/s44172-026-00791-5

Image Credits: AI Generated

DOI: 10.1038/s44172-026-00791-5

Keywords: 6G networks, base-station cooperation, channel feedback, decoupled radio access, resource scheduling, network energy consumption, wireless simulations, communications engineering, prototype testing, Flexible, multi-base-station, cooperation

The study, published in Communications Engineering, was led by Jinchao Liang and Ze Zhang of the Aerospace Information Research Institute at the Chinese Academy of Sciences, together with colleagues from the University of Chinese Academy of Sciences, Aerospace Information University, and partner institutions. Their method, which they call pixel-segmenting imaging, rests on a deceptively simple observation: a pixel does not respond uniformly to light across its area. Quantum efficiency, the probability that an incoming photon generates an electron that the readout circuit can count, varies from point to point within each pixel because of the way microlenses, color filters, and silicon doping are laid down during fabrication. If those variations can be mapped precisely, each pixel becomes a tiny array of detectors with different sensitivities, and the raw sensor output becomes a mixture that can be mathematically unscrambled.

Unscrambling that mixture requires knowing the sensitivity map with high accuracy, and this is where the team’s custom optical system and evolutionary algorithms come in. The researchers illuminate the sensor with carefully designed optical patterns and measure how each pixel responds, then use an evolutionary optimization procedure to infer the intra-pixel quantum efficiency distribution. Evolutionary algorithms, which mimic natural selection by breeding candidate solutions and keeping the fittest, are well suited to this kind of inverse problem because the search space of possible sensitivity maps is enormous, nonlinear, and resistant to conventional gradient-based fitting. By iterating through generations of candidate maps and refining them against measured responses, the algorithm converges on a physical characterization of the sensor that would be extremely difficult to obtain by direct measurement alone.

Once the intra-pixel sensitivity map is known, the imaging model changes fundamentally. The light falling on a single pixel is no longer summed into one value; instead, the measured output is treated as a weighted combination of the intensities reaching each of the sixteen sub-regions, with the weights given by the measured quantum efficiencies. With a custom optical system that projects appropriately structured illumination onto the sensor, the team can solve for those sub-region intensities, effectively reconstructing the scene at a resolution sixteen times finer in area than the native pixel grid. The minimal sampling unit drops from one pixel to one-sixteenth of a pixel, which means the sensor can recover high-frequency spatial information that the original images, under conventional sampling, could not contain.

The performance numbers reported in the paper are striking. At a cut-off frequency of 25 line pairs per millimetre, a standard benchmark for imaging sharpness, the modulation transfer function of the reconstructed images reaches 2.3 times that of the original images. The structural similarity index measure, a metric that captures how faithfully an image preserves perceptually meaningful structure, improves by a factor of 3.1 relative to the originals. These are not marginal gains; they represent a substantial recovery of information that conventional sampling theory would suggest is unrecoverable from a given sensor.

What makes the result particularly notable from an engineering standpoint is that it requires no modification of the sensor hardware itself. The pixel-segmenting approach works with the original device structure intact, relying on a custom optical system in front of the sensor and computational reconstruction behind it. This stands in contrast to strategies that shrink pixels, stack them in three dimensions, or add micromirror arrays, all of which demand expensive changes to fabrication processes. Because the method characterizes the sensor as it is, it could in principle be applied to existing cameras, turning a software-and-optics upgrade into a resolution upgrade.

The distinction between this approach and deep-learning super-resolution is worth emphasizing. Neural-network upscalers are trained on large datasets and learn statistical priors about what natural images look like; when they add detail, they are generating plausible content rather than measuring it. The pixel-segmenting method, by contrast, derives its extra resolution from a physical property of the sensor, the spatial variation of quantum efficiency within each pixel, that is measured and verified for the specific device in use. The recovered high-frequency information is therefore grounded in actual photon measurements rather than in learned expectations, which matters for scientific and security applications where fabricated detail is worse than no detail at all.

The implications reach across several domains. In remote sensing, where satellite and aerial cameras are weight- and power-constrained and cannot simply carry larger sensors, squeezing more resolution out of existing pixels could sharpen Earth-observation imagery without new hardware. In microscopy and industrial inspection, where the finest resolvable feature often determines what can be detected, a sixteen-fold increase in effective sampling density could reveal structures previously blurred away. Consumer imaging could benefit as well, since smartphone cameras are already pushing against the physical limits of pixel miniaturization; a computational layer that extracts sub-pixel information would let manufacturers improve sharpness without the penalty of smaller, noisier pixels.

There are, of course, practical considerations that will shape adoption. The method depends on an accurate intra-pixel quantum efficiency map, which must be measured for each sensor or sensor type, and the evolutionary optimization procedure requires computational resources and careful optical calibration. The custom optical system that structures the illumination must also be integrated into the imaging path, which may be straightforward for dedicated instruments but more challenging for mass-market devices. How the approach performs under real-world conditions, with varying illumination, motion, and noise, will determine whether the laboratory gains translate into field performance. The published work demonstrates the principle and quantifies the improvement under controlled conditions, providing the foundation on which such engineering questions can now be addressed.

For a field that has largely assumed the pixel is an indivisible unit of measurement, the study reframes the problem in a way that is likely to attract attention well beyond optical engineering. By showing that the sampling constraint is not the pixel boundary but the uniformity assumption within it, the researchers have opened a path that combines precision optical characterization, evolutionary computation, and image reconstruction into a single pipeline. If the approach proves robust across sensor types and imaging conditions, the humble pixel, the workhorse of digital imaging for half a century, may turn out to have been hiding sixteen detectors in plain sight all along.

Subject of Research: Sub-pixel imaging using intra-pixel quantum efficiency measurement to exceed pixel-level sampling limits in digital image sensors

Article Title: Pixel-segmenting imaging physically exceeds pixel-level sampling constraints via intra-pixel quantum-efficiency measurement enabled by evolutionary algorithms

Article References: Liang, J., Xue, H., Zhang, C., Guan, M., Xiao, H., Sun, C., Liu, Y., Liu, X., Gao, L., & Zhang, Z. (2026). Pixel-segmenting imaging physically exceeds pixel-level sampling constraints via intra-pixel quantum-efficiency measurement enabled by evolutionary algorithms. Communications Engineering. https://doi.org/10.1038/s44172-026-00773-7

Image Credits: AI Generated

DOI: 10.1038/s44172-026-00773-7

Keywords: digital image sensors, super-resolution, quantum efficiency, evolutionary algorithms, pixel-segmenting imaging, modulation transfer function, optical engineering, computational imaging, sampling constraints, high-frequency recovery, Communications Engineering, image reconstruction

The work, published in Communications Engineering by a collaboration spanning the Toyota Research Institute of North America, the University of Southern California, and Northwestern University, combines two ideas that rarely appear together: a nineteenth-century acoustic device known as a Helmholtz resonator, and a modern neural network trained to translate sound into voltage waveforms. The result is a proof-of-concept sensor system that listens to the heart in the 50 to 120 hertz frequency band and reconstructs the timing of each cardiac cycle with clinically meaningful accuracy.

A Helmholtz resonator is deceptively simple. It consists of a cavity of air connected to the outside world through a narrow neck — the physics behind the tone produced when you blow across the mouth of an empty bottle. The volume of air in the neck acts as a mass and the air in the cavity acts as a spring, and together they oscillate at a characteristic resonant frequency determined by the geometry of the neck and cavity. By carefully designing those dimensions, the researchers engineered a cavity that resonates precisely within the low-frequency band where cardiac acoustic activity concentrates. At resonance, the device amplifies the faint pressure fluctuations produced by the beating heart, boosting a signal that would otherwise drown in ambient noise and in the sensitivity limits of conventional microphones.

This passive amplification is what distinguishes the approach from earlier attempts at acoustic heart sensing. Microphones and accelerometers can pick up cardiac sounds, but the signals of interest sit at extremely low frequencies, where electronic noise, motion artifacts, and environmental rumble dominate. A resonant cavity tuned to the cardiac band acts as a mechanical filter and amplifier in one, emphasizing the physiological content before the signal ever reaches an electronic transducer. Because the amplification is acoustic rather than electronic, it adds no noise of its own and requires no power — a meaningful advantage for any future device that must operate unobtrusively for hours at a time.

Collecting the sound, however, is only half the challenge. The acoustic signal contains rich information about the timing of cardiac events, but it does not look like an electrocardiogram. To bridge that gap, the team developed a neural network-based machine learning algorithm that learns the mapping between the resonator-enhanced acoustic signal and the simultaneously recorded electrical activity of the heart. The network is trained on paired data — acoustic recordings and reference ECG waveforms acquired at the same time — and learns to reconstruct the characteristic features of the cardiac electrical cycle from the acoustic input, including the timing of the R-peak, the sharp spike that marks ventricular depolarization and anchors nearly every heart-rate analysis.

The performance figures reported in the study demonstrate that this translation is more than a curiosity. Compared against a simultaneously acquired reference ECG, the system estimated heart rate with an error of 4.23 plus or minus 4.50 beats per minute, and it located R-peaks with a mean absolute timing error of 36.50 plus or minus 53.17 milliseconds. For context, the two quantities most commonly extracted from ambulatory cardiac monitoring — heart rate, the average beat-to-beat value over time, and heart rate variability, the fluctuation of beat-to-beat intervals — depend directly on how accurately the inter-beat interval can be measured. An error of a few dozen milliseconds in R-peak timing is small enough to support meaningful inter-beat interval determination, which is precisely what the team set out to establish.

Why go to such lengths to recreate an ECG from sound when reliable electrode-based monitors already exist? The answer lies in the friction of everyday health monitoring. Standard ECG measurement requires galvanic contact with the skin, typically through adhesive electrodes or the conductive surfaces of a wrist-worn device. That contact introduces practical constraints: it demands correct placement, adequate skin preparation, sustained pressure, and user compliance. Contact-based sensing dominates today’s wearable landscape, but it is awkward in situations where the user is seated, clothed, or engaged in activities that make electrode contact unreliable. A sensor that works entirely through sound could in principle be embedded into a seat, a headrest, or a cabin structure, monitoring cardiac activity without the wearer ever putting anything on.

That use case is not incidental to the research. The authors explicitly point toward the potential extension of the hardware-plus-software system for non-invasive cardiac biomonitoring of seated users in a dynamic vehicle environment. A car cabin is, in this framing, a plausible sensing platform: the occupant sits in a fixed position for the duration of a trip, and surfaces in contact with or near the body could host an acoustic resonator tuned to the cardiac band. In such a scenario, the vehicle itself would continuously estimate the occupant’s heart rate and heart rate variability, potentially flagging stress, fatigue, or emerging cardiac events during driving. The team’s funding disclosure supports this reading — the experimental work was supported in part by MIRISE Technologies, a supplier of in-cabin sensing technologies for the automotive industry.

The commercial trajectory of the idea is already visible in the intellectual property record. Two of the authors hold a granted United States patent titled Vehicle heart rate detection device, resonators, and methods of detecting a heart rate, and a further patent application covering systems and methods for noncontact monitoring of cardiac activities lists several of the same researchers. The published paper itself, meanwhile, is framed deliberately as a design, experimental testing, and proof-of-concept validation — the authors are careful about what has and has not been demonstrated. The measurements were validated against reference ECGs in controlled conditions; translating that performance to the noise, vibration, and motion of a real vehicle remains future work, as the authors themselves discuss.

Even so, the conceptual advance is striking. It suggests a division of labor between physics and machine learning that could reshape how physiological signals are harvested: let a carefully engineered passive structure do the work of isolating and amplifying the signal of interest, then let a trained network do the work of translating that signal into the clinical vocabulary of medicine. The heart has always been audible if you knew how to listen. What this research demonstrates is that, with the right cavity and the right algorithm, the listening can be precise enough to recover the electrical rhythm itself — beat by beat, without touching the patient.

Subject of Research: Non-invasive cardiac sensing using an acoustic Helmholtz resonator and machine learning-based ECG waveform reconstruction

Article Title: Non-invasive cardiac sensing via an acoustic Helmholtz resonator cavity with electrocardiogram waveform reconstruction

Article References: Schmalenberg, P. D., Hamidi Shishavan, H., Singh, P., Avramidis, K., Severgnini, F. M. Q., Pardo, B., Lee, T., & Dede, E. M. (2026). Non-invasive cardiac sensing via an acoustic Helmholtz resonator cavity with electrocardiogram waveform reconstruction. Communications Engineering. https://doi.org/10.1038/s44172-026-00792-4

Image Credits: AI Generated

DOI: 10.1038/s44172-026-00792-4

Keywords: cardiac sensing, Helmholtz resonator, electrocardiogram, machine learning, neural network, heart rate variability, non-invasive monitoring, acoustics, biomonitoring, in-ear sensing, wearable health, Communications Engineering

The benchmark, called OpenSeeSimE, consists of more than 200,000 question-answer pairs spanning 10,000 parametrically varied simulations. According to the authors, Jessica Ezemba, Jason Pohl, Conrad Tucker, and Christopher McComb of Carnegie Mellon’s Department of Mechanical Engineering, the dataset represents an 850-fold scale increase over what had previously been possible in this niche, enabling statistically robust evaluation across diverse simulation configurations and question types. That scale matters enormously. Small benchmarks can produce results that look impressive but dissolve under statistical scrutiny; with hundreds of thousands of question-answer pairs, the researchers could measure model performance with enough statistical power to detect even incremental progress, and to show confidently when there was none.

The motivation behind the work stems from a genuine bottleneck in engineering design cycles. Interpreting simulation outputs, whether from structural analyses, fluid dynamics studies, or thermal models, requires expensive domain expertise. Engineers must validate complex outputs to ensure safety and performance, and that validation step can slow product development considerably. Large language models have been proposed as assistants in this interpretation task, but they face a fundamental scalability limitation: even modest simulations produce data that exceed the context windows of the best-in-class LLMs. A simulation that generates millions of data points simply cannot be fed into a model that processes text sequentially, no matter how large its context window has grown.

Vision-language models offer a promising alternative, and the reasoning behind that promise is elegant. Rather than ingesting raw simulation data, a VLM can process a visualization of the results, a contour plot, a deformed geometry, a field map, as a compressed representation of the underlying information. VLMs have already demonstrated success across technical visual reasoning domains, from medical imaging to materials characterization. If a model can identify anomalies in radiographs or classify microstructures in electron micrographs, why not read a stress distribution from a finite element plot? The problem, the Carnegie Mellon team found, was that nobody had rigorously tested this assumption, largely because large-scale evaluation frameworks did not exist and expert annotation of simulation outputs was prohibitively expensive.

OpenSeeSimE addresses both obstacles through parametric generation. By systematically varying simulation parameters across 10,000 simulations, the researchers could produce an enormous volume of question-answer pairs without relying on scarce expert annotators to label each example individually. The parametric approach also ensures diversity: the benchmark covers a wide range of simulation configurations and question types, so a model cannot succeed by memorizing a handful of visual patterns. The team built the simulation pipeline using PyAnsys, and the authors acknowledge support and guidance from Raunak Borker, Sanjay Ranganayakulu, Chris Hawkins, and the broader Ansys team in utilizing and understanding the software.

The evaluation itself covered ten state-of-the-art vision-language models, and the findings were striking in their consistency. Models that demonstrate strong performance on general visual reasoning benchmarks, the very tasks that have fueled excitement about multimodal AI, scored between 29 and 47 percent on the engineering simulation questions. For context, random chance on multiple-choice questions typically falls in that range, meaning the models were effectively guessing. The effect sizes were negligible, indicating that the differences between models were not meaningful and that none of them had developed even a rudimentary grasp of simulation interpretation.

This disconnect between general visual competence and domain-specific failure carries a broader lesson about the current generation of AI systems. The pattern recognition abilities that allow a VLM to caption a photograph or answer questions about natural images do not automatically transfer to the stylized, information-dense visual language of engineering analysis. Contour plots encode quantitative information through color scales that must be read precisely; meshed geometries convey structural information through conventions that differ sharply from natural scenes. A model trained predominantly on internet images and text has little exposure to these conventions, and the benchmark results suggest that exposure to general visual data provides essentially no foundation for simulation literacy.

The practical implication, according to the authors, is that deploying vision-language models for simulation interpretation will require domain-specific training rather than reliance on general-purpose models. Off-the-shelf multimodal assistants, however impressive their demo performances, cannot be trusted to read a simulation output today. But the benchmark also provides the infrastructure needed to change that. Because OpenSeeSimE is large enough to measure incremental progress with statistical confidence, researchers developing domain-adapted models can use it to track whether fine-tuning and specialized training data are actually working, rather than relying on small evaluation sets that produce noisy, unreliable signals.

The study also establishes critical baselines for the field. Knowing that current models perform at chance levels gives future researchers a clear starting point against which to measure improvement. The benchmark’s reusable framework, with its 850-fold scale increase over prior evaluation efforts, is designed to serve as a lasting resource for the community. As AI systems are increasingly proposed for safety-critical engineering tasks, from validating structural designs to checking thermal performance, rigorous evaluation of their actual capabilities becomes not just scientifically interesting but essential for responsible deployment.

Published open access in Communications Engineering on 23 September 2026, the paper arrives at a moment when the gap between AI hype and AI capability in specialized domains is under intense scrutiny. The Carnegie Mellon work adds engineering simulation to the growing list of technical areas, alongside medicine, law, and scientific research, where general-purpose models stumble without targeted training. For engineers hoping that AI might soon ease the burden of interpreting complex simulation outputs, the message is clear: the tools are not ready yet, but for the first time, there is a rigorous, large-scale yardstick to measure how close they are getting.

Subject of Research: Benchmark evaluation of vision-language models for engineering simulation question answering

Article Title: A large-scale benchmark to assess vision-language model question answering capabilities in engineering simulations

Article References: Ezemba, J., Pohl, J., Tucker, C., & McComb, C. (2026). A large-scale benchmark to assess vision-language model question answering capabilities in engineering simulations. Communications Engineering. https://doi.org/10.1038/s44172-026-00786-2

Image Credits: AI Generated

DOI: 10.1038/s44172-026-00786-2

Keywords: vision-language models, engineering simulations, benchmark, OpenSeeSimE, question answering, machine learning, mechanical engineering, Carnegie Mellon University, visual reasoning, finite element analysis, AI evaluation, Communications Engineering

A new analysis published in Communications Engineering examines how two familiar wave phenomena, reflection and transmission, can be harnessed systematically to keep terahertz links alive when the direct path between transmitter and receiver is blocked. The work argues that non-line-of-sight terahertz communication should not be treated as an afterthought or a failure mode, but as a first-class design objective, engineered deliberately into the architecture of future sixth-generation networks. By treating reflective surfaces such as walls, ceilings, glass facades, and even road signs as passive allies, and by exploiting controlled transmission through materials such as drywall and window glass, the study sketches a pathway toward terahertz coverage that survives the shadows cast by everyday obstacles.

The scale of the challenge is rooted in fundamental electromagnetics. As frequency increases, the free-space path loss between two antennas grows with the square of the frequency, so a terahertz link at 300 gigahertz suffers dramatically more spreading loss than a microwave link carrying the same distance. Diffraction, the phenomenon that allows lower-frequency radio waves to bend around corners and fill rooms with coverage, becomes vanishingly weak at terahertz frequencies because the wavelengths, on the order of a millimeter or less, are tiny compared with the edges and gaps of ordinary objects. Scattering from rough surfaces, which would smear energy in many directions at lower frequencies, becomes more specular and directional, meaning that a terahertz wave bounces off a flat wall much the way a billiard ball bounces off a cushion: predictably, but only along a narrow geometric path determined by the law of reflection.

This directional character is a double-edged sword. On one hand, it means that a blocked terahertz link cannot rely on the diffuse, multipath-rich propagation that made older cellular systems robust. On the other hand, it means that reflected paths are stable, identifiable, and potentially very high quality. A signal that bounces once off a well-positioned wall arrives with a predictable delay, a predictable angle of arrival, and, if the surface is sufficiently smooth and conductive, only a modest additional loss. The new analysis emphasizes that in many indoor and street-canyon scenarios, first- and second-order reflected paths can deliver received power levels well within the sensitivity range of modern terahertz receivers, provided that the antenna beams are steered to illuminate and collect from those paths deliberately.

Beam steering is therefore the linchpin of the approach. Terahertz systems almost universally employ highly directional antennas, often implemented as phased arrays of many small radiating elements, to concentrate their limited power into a narrow pencil beam and to overcome path loss. In a line-of-sight world, pointing such a beam is straightforward. In a non-line-of-sight world, the network must first discover which reflective or transmissive paths exist, then select the best one, and then keep the beam locked onto it as people move and doors swing open. The analysis highlights beam management as a central research problem: hierarchical search procedures, machine-learning-assisted path prediction, and the use of out-of-band or sub-6-gigahertz control channels to cue the terahertz beam are all candidate techniques for finding a viable reflected route in milliseconds rather than seconds.

Reconfigurable intelligent surfaces, sometimes called smart metasurfaces, extend this idea from accidental reflectors to engineered ones. These are thin panels studded with subwavelength elements whose electromagnetic response can be tuned electronically, allowing the panel to act as a programmable mirror that redirects an incoming terahertz beam toward a chosen receiver, or even to shape the wavefront in ways a plain wall never could. Mounted on ceilings, building facades, or interior walls, such surfaces could serve as deliberate relay points, stitching together coverage in courtyards, corridors, and factory floors where no natural reflection geometry happens to work. The study situates these engineered surfaces alongside natural reflections as complementary tools in the non-line-of-sight toolkit, noting that their low power consumption and passive operating principle make them attractive for dense deployment.

Transmission, the second pillar of the analysis, is often dismissed at terahertz frequencies because many common building materials attenuate the signal severely. Concrete and brick are effectively opaque, and even wood and glass introduce losses that grow rapidly with frequency and thickness. Yet the picture is nuanced. Drywall, plasterboard, and standard window glass can pass usable fractions of the incident power at the lower end of the terahertz range, particularly near 100 to 140 gigahertz, where several standards bodies are now defining early commercial allocations. A signal that passes through a single interior wall may arrive attenuated but decodable, especially when the transmitter can boost power or the receiver can integrate over longer symbol durations. The analysis argues that material-aware network planning, in which the propagation database records not only reflectivity but also transmissivity of the surfaces in a given environment, allows the network to choose intelligently between penetrating an obstacle and bouncing around it.

The interplay between reflection and transmission also shapes how channel models must evolve. Classical statistical channel models, built for scattering-rich microwave environments, poorly describe terahertz links, which are dominated by a small number of deterministic geometric paths. Ray-tracing simulations, seeded with accurate three-dimensional models of buildings and streets and with measured material parameters, emerge as the preferred planning tool. The analysis stresses that measurement campaigns at terahertz frequencies remain comparatively sparse, and that closing this gap between modeled and measured channel behavior is essential before operators can trust simulated coverage maps. Hybrid approaches that combine sparse measurements with ray-tracing interpolation, updated in real time from the feedback that connected devices report, offer a pragmatic middle path.

The implications reach well beyond faster smartphone downloads. Terahertz non-line-of-sight techniques are critical for industrial wireless networks, where metal machinery guarantees that direct paths will be blocked, and where deterministic, ultra-reliable low-latency communication is a hard requirement for replacing factory cabling. They matter for vehicular networks, where the geometry of a street canyon creates rich reflection opportunities off building walls and parked vehicles, and for integrated sensing and communication, in which the same reflected terahertz signals that carry data can simultaneously image the environment. They also underpin the vision of wireless data centers and chip-to-chip links, where racks and heat sinks create a maze of reflective metal surfaces that a well-designed terahertz system can exploit rather than suffer.

Considerable engineering hurdles remain before reflected and transmitted terahertz paths can be managed at network scale. Phased arrays with thousands of elements consume significant power, and the phase shifters and amplifiers behind them are still costly at these frequencies. Beam alignment protocols must become faster and more energy-efficient, and handover between line-of-sight and non-line-of-sight states must be seamless enough that users never notice the network quietly rerouting their connection around a passing pedestrian. Standardization bodies are actively working to define the measurement, signaling, and beam-management procedures that would let multiple vendors implement interoperable non-line-of-sight behavior. What the new analysis makes clear is that the physics need not be an enemy. The same specular, mirror-like behavior that makes terahertz waves unforgiving also makes them predictable, and predictability is precisely what a carefully engineered network needs. If future 6G systems learn to treat every wall and window as potential infrastructure, the terahertz frontier may arrive not despite the obstacles in its path, but because the network has learned to use them.

Subject of Research: Non-line-of-sight terahertz communications using reflection and transmission for future 6G networks

Article Title: Non-line-of-sight terahertz communications enabled by reflection and transmission

Article References: Liu, F., Seddon, J., Belio-Apaolaza, I., G. Huggard, P., Wang, H., Lia, E., Deborgies, F., Wu, J. E., Munshi, M. N., Fice, M., & Seeds, A. (2026). Non-line-of-sight terahertz communications enabled by reflection and transmission. Communications Engineering. https://doi.org/10.1038/s44172-026-00782-6

Image Credits: AI Generated

DOI: 10.1038/s44172-026-00782-6

Keywords: terahertz communications, non-line-of-sight, 6G, beam steering, reconfigurable intelligent surfaces, reflection, transmission, channel modeling, millimeter wave, wireless networks, Communications Engineering, phased arrays

For more than a century, the standard playbook for improving convective heat transfer has been additive and largely empirical. Engineers added fins, ribs, dimples, and vortex generators to surfaces, tested them in wind tunnels and flow loops, and iterated until performance improved. Each geometry was a hypothesis tested forward: build it, measure it, and hope the enhancement outweighed the penalty of added pressure drop, added weight, or added manufacturing cost. The trouble with this forward approach is that the design space is astronomically large. The height, width, spacing, curvature, and arrangement of surface features each influence the boundary layer, the near-wall velocity gradients, and the temperature field in coupled and often nonlinear ways. Searching that space by trial and error is slow, expensive, and rarely converges on a truly optimal geometry.

Reverse design, sometimes called inverse design, flips the logic of the problem. Instead of asking what a given bump shape does to heat transfer, the designer specifies a target, for example a desired local heat flux distribution along a wall, a target Nusselt number profile, or a uniform surface temperature under a fixed thermal load, and then asks what interfacial geometry produces it. Mathematically, this means treating the governing equations of fluid flow and heat transport, the Navier-Stokes equations coupled with the energy equation, as constraints in an optimization problem whose unknowns are the geometric parameters of the surface itself. Gradient-based optimization, adjoint methods, and increasingly machine-learning surrogate models make it feasible to navigate this inverse problem efficiently, even when the mapping from shape to performance is highly nonlinear.

The appeal of bumps, as opposed to extended fins, lies in their compactness and their passive nature. A bump on a heated wall perturbs the flow in a controlled way: it thins the thermal boundary layer at its windward face, accelerates flow over its crest, and can promote periodic flow reattachment and mixing in its wake. Unlike active cooling strategies that demand pumps, fans, or external power, passive enhancement exploits the natural motion of the fluid. This matters enormously in applications where power budgets are tight or reliability is paramount, from aerospace thermal protection to sealed power electronics modules to passive safety systems in next-generation nuclear reactors, where cooling must continue even when all powered systems fail.

The study’s central contribution is a systematic framework that connects the geometry of interfacial bumps to the heat transfer outcome they produce, and then inverts that connection. By parameterizing bump shapes with a flexible set of descriptors, including height, base width, curvature, and streamwise spacing, the researchers map how each parameter influences key performance metrics such as the average Nusselt number, the local heat flux distribution, and the pressure drop penalty. Crucially, the framework distinguishes between enhancement that comes from simply increasing surface area and enhancement that comes from genuinely improving the convective transport at the fluid-solid interface. This distinction matters because area gain is cheap and predictable, while true transport improvement is where the interesting physics lives.

One of the most instructive findings concerns the trade-off between thermal performance and hydraulic resistance. Any protrusion into a flow channel extracts a price in pressure drop, which translates directly into pumping power. A naive design that maximizes heat transfer coefficients often does so by generating turbulence and form drag, and the energy cost of pushing fluid through the channel can erase the thermal benefit. The reverse design approach makes this trade-off explicit and quantitative. By encoding the pressure penalty directly into the objective function, the optimization can search for bump geometries that sit on the Pareto frontier, the set of designs for which no further thermal gain is possible without additional flow resistance. The resulting shapes are often counterintuitive: rather than sharp, tall obstacles, the optimized bumps tend toward smoothly varying profiles that manipulate the near-wall flow gently but persistently.

The physics underlying the optimized geometries is rooted in boundary layer management. When fluid flows over a smooth heated plate, a thermal boundary layer grows along the surface, and because the layer thickens downstream, the local heat transfer coefficient decays. A well-placed bump interrupts this growth, forcing the boundary layer to restart and restoring high local heat transfer coefficients. But the timing and character of that interruption are everything. A bump that is too abrupt separates the flow, creating a recirculation bubble in which heat transfer can actually stagnate or degrade. A bump that is too gentle may barely perturb the layer at all. The inverse framework identifies the intermediate regime, where the bump’s curvature keeps the flow attached while its height and streamwise position are tuned so that each successive feature re-energizes the boundary layer at the optimal moment.

Another dimension of the work is its treatment of spatially nonuniform targets. In many practical devices, heat generation is not uniform: hotspots appear near transistor junctions, fuel rod spacers, or laser diode facets. Uniform enhancement wastes cooling capacity where it is not needed. The reverse design framework can specify a nonuniform target heat flux or temperature profile and solve for a bump array whose geometry varies along the surface to deliver cooling exactly where it is required. This capability points toward a future of thermally tailored surfaces, in which the microgeometry of a wall is designed jointly with the device it cools, rather than as an afterthought. Combined with additive manufacturing, which can now fabricate complex three-dimensional surface textures at fine scales in metals and polymers, such tailored surfaces are moving from concept to manufacturable reality.

The methodology also highlights the growing role of computational tools in thermal engineering. Solving inverse problems requires many evaluations of the forward physics, typically through computational fluid dynamics simulations that resolve the flow and temperature fields around each candidate geometry. Modern adjoint techniques reduce the cost of computing sensitivity information, allowing gradient-based optimizers to converge in far fewer iterations than brute-force searches. Where full simulations remain too expensive, reduced-order models and machine-learning surrogates trained on simulation data can stand in during the search, with high-fidelity checks applied to the final candidates. The study’s framework is consistent with this broader trend: thermal design is becoming an optimization-driven discipline in which intuition guides the setup of the problem, but computation explores the design space.

The implications extend across a remarkable range of scales and sectors. In data centers, where cooling already consumes a significant fraction of total energy use, passively enhanced surfaces in cold plates and immersion systems could reduce parasitic power draw. In electronics packaging, hotspot-targeted bump arrays could extend device lifetimes by flattening temperature gradients that drive thermomechanical fatigue. In energy conversion, from gas turbine blade cooling to heat exchangers in concentrated solar plants, even single-digit percentage improvements in convective performance translate into substantial efficiency and cost gains at scale. And in passive safety systems, where cooling must function without any external power, geometries that squeeze more performance out of natural convection carry a direct safety dividend.

Of course, the path from optimized simulation to robust hardware involves challenges that the study’s framework does not by itself resolve. Real surfaces foul, erode, and corrode; manufacturing tolerances blur the fine distinctions between nominally optimal shapes; and flow conditions in service vary beyond the design point, so a geometry optimized for one Reynolds number may underperform at another. The authors’ contribution is best understood as establishing a principled starting point: a rigorous, invertible relationship between bump geometry and heat transfer behavior that engineers can adapt, constrain, and validate experimentally. As fabrication technologies mature and computational design tools become routine, the idea that a surface’s microscopic landscape can be computed rather than guessed represents a quiet but meaningful shift in thermal engineering, one that turns the humble bump from a rule-of-thumb feature into a precisely specified instrument of passive heat transfer control.

Subject of Research: Reverse design of fluid-solid interfacial bump geometries for passive convective heat transfer enhancement

Article Title: Reverse design of fluid-solid interfacial bumps for passive heat transfer enhancement

Article References: Yang, G., Zhou, X., Liu, X., Xiong, J., Chai, X., Zhang, T., & He, H. (2026). Reverse design of fluid-solid interfacial bumps for passive heat transfer enhancement. Communications Engineering. https://doi.org/10.1038/s44172-026-00772-8

Image Credits: AI Generated

DOI: 10.1038/s44172-026-00772-8

Keywords: heat transfer, reverse design, inverse optimization, fluid-solid interface, passive cooling, boundary layer, Nusselt number, pressure drop, thermal management, computational fluid dynamics, additive manufacturing, Communications Engineering