The FNR is pleased to announce that 5 projects have been retained for funding in the 2025-2 BRIDGES Call, an FNR commitment of 1.69 MEUR. The BRIDGES programme provides financial support for industry partnerships between public research institutions in Luxembourg and national or international companies. The next deadline is 29 September 2026, 14:00 CET.
Funded projects
| PI | Coordinating Institution | Private Partner | Project Title | Acronym | Domain | FNR funding |
| Stephan Leyer | University of Luxembourg | Meerstetter Engineering GmbH | Functional Advanced Composites For Thermoelectric Systems | FACTS | Physics, Engineering & Material Sciences | € 130,000.00 |
| Abstract | Across electronics and energy systems, precise heat-flux measurement still leans on thermopile designs that are thick, slow, and costly to scale, and the field has lacked a process-integrated route that combines high sensitivity with traceable calibration and reproducibility. FACTS addresses this gap by advancing functional advanced composites built as architected multilayers for the transverse thermoelectric effect (TTE) and by formalising a model-to-metrology workflow aligned with standard thin-film tooling and established manufacturing practice. The project builds directly on SELENA, which delivered the first reproducible multilayer route for this sensor class, stood up the end-to-end build flow, and put in place mirrored cross-site calibration, run cards, and uncertainty analysis. That foundation created a pilot-ready path and a common toolchain across partners, allowing FACTS to focus on performance gains rather than recreating basic infrastructure. Two technical strands run in parallel: simulation-guided down-selection introduces process-compatible material substitutions inside the proven multilayer architecture to raise the effective transverse coefficient while preserving throughput, cost, and safety; in tandem, voltage-boosting device layouts increase effective path length without compromising the thermal path or response time. Both strands operate within an uncertainty-aware methodology with mirrored calibration and defined process windows, supporting clean transfer from lab to pilot production. The programme targets one to two orders of magnitude improvement in device-level sensitivity over the established baseline, with projections grounded in extensive prior numerical work that identified capable candidate materials and quantified gains under matched boundary conditions. These improvements will be verified under traceable calibration with combined uncertainty at or below 10 percent, and delivered as calibrated pre-series units alongside an analytical framework that standardises figures of merit and comparison across laboratories, including uncertainty methods, calibration workflows, and process documentation for replication and scale-up. Across electronics thermal control, battery management, and materials characterisation, compact calibrated transverse sensors can tighten control loops, shorten test cycles, and enable energy-aware process decisions; the same composites-and-architecture approach also opens a credible path toward practical waste-heat harvesting and transverse Peltier devices using industry-friendly materials and flows. | |||||
| PI | Coordinating Institution | Private Partner | Project Title | Acronym | Domain | FNR funding |
| Facundo Bre | Luxembourg Institute of Science & Technology (LIST) | Paul Wurth Geprolux (PWGP) | Smart Planning Renovation Insights For A Decarbonised Built Environment | SPRINT | Sustainable resources | € 398,000.00 |
| Abstract | Renovating existing buildings is essential to achieving climate neutrality in Europe, yet decision-makers still lack robust and transparent tools to plan effective interventions. The SPRINT project (Smart Planning Renovation INsights for a decarbonised builT environment) addresses this challenge by developing a digital decision-support platform for the renovation of public buildings in Luxembourg. SPRINT integrates three complementary dimensions of building performance into a single workflow: energy simulation, life cycle assessment (LCA), and life cycle costing (LCC). These components are combined in a multi-objective optimisation framework that allows stakeholders to identify renovation strategies balancing environmental performance, energy savings, and economic feasibility. A key methodological innovation lies in the use of machine learning surrogate modelling, which accelerates detailed energy simulations. The whole automated pipeline aims to reduce analysis time from months to weeks while maintaining accuracy. The project also introduces marginal abatement cost curves (MACCs) into building renovation planning. By expressing greenhouse gas reductions relative to financial investment, MACCs provide transparent benchmarks for prioritising interventions under constrained budgets. Together with a structured renovation measure database, this approach ensures consistent evaluation of renovation options across different building typologies. Validation will be carried out through 2–3 actual case studies with Luxembourg municipalities, ensuring that the tool reflects real-world building conditions, regulatory requirements, and stakeholder needs. The platform will be delivered through a web-based interactive dashboard, enabling planners, municipalities, and building owners to configure scenarios, explore trade-offs, and compare optimal pathways to decarbonisation. SPRINT creates value at multiple levels. For municipalities and building owners, it accelerates the preparation of renovation plans, provides transparent performance indicators, and supports evidence-based investment decisions. For the industrial partner, it supports the improvement of the consulting service delivered to the clients and strengthens competitiveness in sustainable construction. For the scientific community, it advances the integration of energy simulation, LCA, and LCC, while providing new datasets and methodologies for renovation research. By combining scientific excellence with industrial experience, SPRINT directly supports Luxembourg’s Long-Term Renovation Strategy and National Energy and Climate Plan while contributing to the EU Renovation Wave. The project advances Luxembourg’s leadership in sustainable construction and offers a scalable model for renovation strategy planning across Europe’s decarbonised built environment. | |||||
| PI | Coordinating Institution | Private Partner | Project Title | Acronym | Domain | FNR funding |
| Mahesh Desai | Luxembourg Institute of Health (LIH) | MEDICE Arzneimittel Pütter GmbH & Co. KG, Germany | Innovative Strategies To Inform Customized Gut Health Treatment | INSIGHT | Life Sciences, Biology & Medicine | € 400,000.00 |
| Abstract | The INSIGHT project builds on our previous FNR BRIDGES project (B-MARK, 22/17426243) through the development and validation of the MEDIBIOM Gut Health Test (GHT), paralleled by evaluation of a postbiotic intervention for patients with IBS-like disorders. The project integrates nine work packages: WP A (A1–A4) focuses on assay development and validation of microbial, enzymatic, and host biomarkers, while WP B (B1–B4) establishes a patient cohort to test and refine personalized nutritional treatments using MEDIBIOM (part of MEDICE) products. Data from both arms will feed into predictive models using machine learning, enabling individualized treatment recommendations. Oversight and integration in WP C ensure scientific excellence, IP protection, and dissemination. INSIGHT will deliver validated diagnostics, personalized interventions, and translational pathways that support MEDICE’s strategy of advancing holistic, evidence-based healthcare solutions. | |||||
| PI | Coordinating Institution | Private Partner | Project Title | Acronym | Domain | FNR funding |
| Eleni Koronaki | Luxembourg Institute of Science & Technology (LIST) | Invitrolize | AI For Respiratory Sensitization Assessment And Fair Evaluation | AIRSAFE | Innovation in services | € 398,000.00 |
| Abstract | AIRSAFE is a collaboration between Invitrolize and the Luxembourg Institute of Science and Technology (LIST) to transform Invitrolize’s patented ALIsens® assay into a next-generation platform for chemical safety testing. ALIsens® is a validated in vitro model replicating key interactions in the human respiratory system and offering an ethical alternative to animal testing. However, broader industrial and regulatory uptake is limited by three bottlenecks: (i) defining baseline concentrations (CV₇₅) requires labor-intensive testing, (ii) current biomarker panels are large, costly, and hard to interpret, not to mention, highly correlated, and (iii) assay data is fragmented, slowing reporting and integration. AIRSAFE tackles these challenges by embedding AI and data management into Invitrolize’s workflow. A hybrid machine learning–Bayesian framework will predict CV₇₅ with up to 10% error, cutting concentration testing by 30–50%. Interpretable ML will reduce the >25-analyte biomarker panel by half while retaining ≥90% accuracy, enabling reliable multi-class classification of sensitizers, irritants, and non-reactives. A FAIR-compliant data and reporting system will harmonize results, protect proprietary information, and use a large language model to draft reports, reducing turnaround by up to 50%. The result will be a faster, more cost-efficient, and transparent service that boosts Invitrolize’s competitiveness in the expanding market for New Approach Methodologies (NAMs). Beyond commercial benefits, AIRSAFE advances the EU’s 3Rs principles, supporting the global shift toward ethical, human-relevant toxicology and reinforcing Luxembourg’s position in data-driven life sciences innovation. | |||||
| PI | Coordinating Institution | Private Partner | Project Title | Acronym | Domain | FNR funding |
| Djamila Aouada | University of Luxembourg | LMO | Unsupervised Domain Adaptation For Robust Spacecraft Pose Estimation Via Regression Using Geometric Representations And Uncertainty-aware Mechanisms | U-ADAPT | Innovation in services | € 372,000.00 |
| Abstract | Due to the exponential increase in spacecraft and space debris in orbit, achieving autonomy in orbital proximity operations is becoming critical. This can only be accomplished by accurately estimating the pose of target space objects. Currently, Spacecraft Pose Estimation (SPE) methods relying on monocular RGB sensors are being widely explored due to their compatibility with small satellites. However, these approaches are mostly based on Deep Learning (DL) models, which often suffer from limited generalization. This challenge is amplified by the reliance on simulated data for training, caused by the scarcity of labelled real-world space data, leading to SPE models that fail to generalize under realistic space conditions. U-ADAPT aims to explore unsupervised domain adaptation to enhance the applicability of SPE methods in real conditions. By doing so, U-ADAPT will advance the state of the art in autonomous proximity operations while reinforcing Luxembourg’s leadership in space innovation and contributing to long-term space sustainability. | |||||