Tech & AI

A systematic review of 175 full-scale operations reveals that centrifuges remove more waste solids and phosphorus from biogas digestate than screw presses, yet consume 4.5 times more energy. For operators weighing efficiency against operating costs, screw presses deliver better nutrient recovery per unit of electricity used.EN

Universities can now train biomedical engineers in electrical safety through interactive remote laboratories, a three-year study shows. The finding matters for hospital networks and educational institutions seeking cost-effective compliance training that maintains engagement across geographically dispersed workforces.EN

A new cost analysis of Sweden's agricultural biomethane potential reveals that a hybrid approach—decentralized biogas production fed through pipelines to centralized liquefaction facilities—often beats fully centralized or fully decentralized setups. The finding could reshape European investments in renewable energy infrastructure as countries race to meet climate targets and reduce fossil fuel imports.EN

A study of language teachers found that classroom response systems—where students answer questions anonymously via their own devices—significantly increased engagement among introverted learners reluctant to speak up. For educational institutions and EdTech vendors, the finding suggests mobile-based polling tools could be a scalable way to improve participation rates across diverse student populations.EN

Researchers testing engineered timber (CLT) discovered that its shear strength behaves very differently under vibration than in static tests—jumping 40% higher dynamically. This matters for architects and builders: existing design codes may underestimate or mispredict how CLT performs in real-world conditions like wind, earthquakes, or vibration, potentially affecting safety margins and material costs.EN

Researchers have cracked a major cost problem in autonomous driving: creating detailed lane maps. Using only basic cameras and GPS already in consumer vehicles, a new system can automatically generate high-precision road data. This could slash the enormous expense of hand-mapping streets for self-driving cars.EN

Scientists have demonstrated a backup navigation system using digital television broadcasts combined with radio ranging, achieving accuracy within tens of meters without relying on GPS. The approach could secure critical infrastructure and autonomous systems against growing threats to satellite navigation networks.EN

Researchers have developed a system using underground sensors to detect and track elephants before they raid villages, achieving 89% accuracy in identifying elephant footsteps from up to 30 meters away. The technology offers a low-cost early warning solution for regions where human-wildlife conflict threatens both rural livelihoods and endangered elephant populations.EN

Researchers have developed an optimization technique that automatically designs internal cooling channels for high-temperature engines—potentially cutting design time and improving efficiency. The approach could accelerate development of more durable turbines for aerospace and power generation, where thermal performance directly impacts operational lifespan and fuel consumption.EN

A new study of Nordic logging operations reveals that adding a third operator to two-person harvesting crews cuts wood delivery delays by half, despite raising per-unit costs. For timber suppliers competing on speed and reliability, the trade-off may be worth it—especially when customer flexibility matters more than squeezing every penny from production.EN

A new paper argues that AI research across industries is stuck optimizing for academic prestige rather than solving actual problems. Researchers propose shifting from competition-based metrics to collaborative frameworks that measure societal impact—a change that could reshape how companies and institutions fund and evaluate AI initiatives.EN

Researchers discovered that engineered metal-dielectric nanostructures can sense temperature changes by measuring how they absorb light rather than emit it. This finding could enable new thermal sensors for computing systems and infrared detection devices that operate more efficiently than existing alternatives.EN

Researchers found that standard laboratory cultures of brain cells fail to capture how the body actually fights tick-borne encephalitis virus in living brains. The gap between test-tube and real-world immune responses could explain why some experimental treatments work in labs but flop in patients.EN

Researchers have identified how a bacterial enzyme self-regulates to punch holes in cell walls, enabling the transfer of antibiotic resistance genes between organisms. The discovery could help pharmaceutical companies design new antibiotics or block resistance spread—a growing threat that costs healthcare systems billions annually.EN

Researchers have engineered stable semiconductors with tunable magnetic properties that could power future spintronic devices and data storage systems. The breakthrough uses lead-free materials that are cheaper and safer to manufacture than current alternatives, potentially opening a new market for sustainable electronics manufacturers.EN

Researchers have cracked a foundational problem in network design: determining how densely connected a network must be to guarantee it contains optimal routing paths. The findings could improve everything from telecommunications infrastructure to supply chain logistics by clarifying minimum connectivity requirements.EN

Researchers have identified four design principles for teaching hypothesis testing to 14-16 year-olds using hands-on experiments and digital simulations. The approach—tested by having students taste-test cola brands—could reshape how schools teach statistical reasoning, a skill increasingly vital for data-driven decision-making across industries.EN

Researchers validated a flexible smart patch that reliably measures running step rates, matching or exceeding the accuracy of established fitness trackers and lab equipment. The finding opens a path for seamless health monitoring embedded directly in clothing—a potential market opportunity for wearable manufacturers seeking competitive advantages in athletic performance tracking.EN

A new analysis reveals that education policy guidelines for AI adoption contain competing visions of learning itself—some emphasizing personalization, others efficiency or equity. These conflicting frameworks risk creating fragmented systems that don't deliver on promises, with profound implications for how vendors design products and governments allocate education budgets.EN

Scientists have more than doubled the electrical voltage generated from waste heat by mimicking how human skin senses temperature. The advance could unlock new applications for thermoelectric devices in industrial waste recovery and wearable electronics, where capturing ambient heat as usable power remains commercially valuable.EN

Deep learning systems are stepping in to analyze the explosion of medical images—CT scans, MRIs, X-rays—that overwhelm radiologists and delay diagnoses. As hospitals digitize records and expand screening programs, AI offers a path to faster cancer detection while freeing clinicians from repetitive image review.EN

Researchers have developed a machine learning system that forecasts where people will move in complex environments by mimicking how air flows around obstacles. The advance could improve safety systems for autonomous vehicles and robots operating near pedestrians, addressing a critical gap in current prediction technology.EN

Researchers have developed a learning system that enables robots to push, rotate, and reposition objects using their environment—even when they can't see how to grasp them directly. The breakthrough could unlock robotic automation for warehouses and manufacturing lines handling items in awkward positions or cluttered spaces.EN

Researchers have developed a more efficient neural network method for converting camera images into detailed 3D surface maps—cutting memory requirements to just 10-50% of existing approaches. The breakthrough matters for robotics companies, autonomous systems, and any operation needing to reconstruct environments from sparse visual data without breaking hardware budgets.EN

Adverse weather cuts object detection accuracy in autonomous systems, creating safety and liability risks. Researchers tested three approaches—real-world training data, synthetic weather simulation, and image denoising—to restore reliability. The findings could accelerate deployment of self-driving technology in regions with harsh winters.EN