Skip to content
About FNR
Funding
Funding process & policies
News, Media & Events

Luxembourg National Research Fund

Results Industrial Fellowship 2022-1 Call

The FNR is pleased to communicate that 7 of 11 Industrial Fellowship projects (6 PhD, 1 Postdoc) have been selected for funding in the 2022-1 Call, representing an FNR commitment of 1.35 MEUR. 

The aim of the Industrial Fellowships programme is to foster the cooperation between Luxembourg based companies active in R&D and public research institutions in Luxembourg and/or abroad. The scheme awards PhD and Postdoc grants to researchers who carry out their PhD and/or postdoc training in collaboration with a company in Luxembourg. The scheme is open to all scientific domains, and to all researchers, regardless of their nationality. Collaborating companies must have a presence in Luxembourg.

Go to Industrial Fellowships programme page – 2022-2 Call deadline: 6 October 2022, 14:00 CET.

Funded Industrial Fellowships projects

PhD – ICT & Space

Applicant

Elona Dupont

Project title

Constrained Sequence modelling of CAD for reverse Engineering from 3d Scans (CASCADES)

Host institution

University of Luxembourg (SnT)

Collaborating company

Artec3D

Abstract

In recent years, Artificial Intelligence (AI) has seen some incredible progress at completing tasks that were thought to be only possible by humans such as speech recognition, sentiment analysis, and even producing visual art. However, AI models still struggle to capture complex tasks that are constrained by different human and technical parameters. One example of such a task is Computer-Aided Design (CAD) that is a technical process driven by human intuition. It can take skilled engineers years to master CAD modelling as it requires learning a sound combination of design and technical skills. By gaining experience, the engineers develop a design intuition which allows them to think ahead and take the right actions at all stages of the design process.

CASCADES will use cutting-edge AI technology to make machines learn this design intent. In particular, technical constraints will be used as clues to uncover the thought processes guiding CAD designers. Such technology could be directly applied to reverse engineer the CAD construction history of a physical object, thus allowing to save resources by automating industrial design. The interest in this reverse engineering process has been encouraged by recent advances in 3D handheld scanner technology.

As a result, the ultimate goal of CASCADES is to develop a complete pipeline to automate the 3D reverse engineering of CAD models from 3D scans by mimicking the design intent. CASCADES will be conducted in close collaboration with one of the world leaders in the field of 3D technologies based in Luxembourg, Artec3D. By automatically learning the design intent, it is hoped that CASCADES will bring us one step closer towards modelling the human mind using machines.

Applicant

Rodrigo da Silva Gesser

Project title

Multi-Layer Model Predictive Control for Urban Water Management (MoCaMa)

Host institution

University of Luxembourg (SnT)

Collaborating institution

RTC4Water

Nowadays, governments are investing heavily to reduce the impact of society over the environment and, with the combination of investments in civil infrastructures and new technologies, these impacts are being constantly diminished. One example of such combination is the construction of urban drainage systems (UDS), responsible for collection, distribution and disposal of wastewater, coupled with advanced intelligent systems that guarantee and improve the functioning of the UDS. These intelligent systems are commonly called Real Time Control (RTC) systems, since they are responsible for constantly monitoring and acting on the wastewater collection and distribution. The greatest challenge in anUDS is the discharge of overflown wastewater to rivers and seas. This is mainly caused by intense rain events that flood the UDS with rainwater, greatly increasing the volume of wastewater in the UDS, resulting in overflows.

Studies show that the use of RTC in urban drainage system reduces the overflowed wastewater and, therefore, contribute to cleaner and safer disposal of waste. The usual implementation of RTC systems does not consider the actual concentration of the pollutant in the wastewater, but this information is beneficial because the RTC system is then able to know that a region is more polluted than other and act to avoid overflows in this polluted area.

Hence, this project proposes a novel RTC system that uses the actual value or an estimative of concentration to improve the overall performance of the UDS, focusing on reducing the concentration on overflowed wastewater. In order to change the current RTC, the new approach increases the complexity of the structures, which impose challenges for controlling the urban drainage system, such not being able to calculate the correct action due to long computational time.

Therefore, it is proposed the inclusion an algorithm that simplifies the RTC system in case of issues, guaranteeing that the RTC is able to always take safe actions. The main objective of this project is, therefore, to develop a control solution that minimizes the pollution discharged in the environment by the UDS while assuring a proper behaviour of the system.

Applicant

Bernhard Specht

Project title

A Hybrid Brain-Computer Interface System for Remote Monitoring of Multiple Sclerosis (MMS)

Host institution

Manchester Metropolitan University

Collaborating company

Myelin-H

Abstract

Recent statistics have shown that more than 2.8 million people suffer from multiple sclerosis (MS) and the disease causes a reduction in life expectancy of 7–14 years. Given that MS is an incurable and silently progressive disease, we have witnessed, over the last decade, the extensive use of different treatments to alleviate the devastating disease’s symptoms. Hence, it has become essential to monitor their effectiveness. Similarly, it has recently become imperative to continuously (24/7) monitor the disease’s progression and diagnose its exact course.

So far, magnetic resonance imaging (MRI) scans remain the sole way to diagnose MS and monitor its progression. However, MRIs are expensive, inaccessible daily, cannot be used at home, and more importantly, lack the temporal changes (compared to electroencephalography (EEG)) to detect millisecond changes in the brain before and after treatments en route to assessing their effectiveness. Notwithstanding the recent endeavours that have been made in developing different digital apps to monitor MS, they have proven their inefficiency to provide remote, real-time, and accurate monitoring of patients’ brain activity before and after medication. This has therefore spurred the use of individual bio signals, specifically EEG or electromyography (EMG) to diagnose and monitor the disease. However, to the best of our knowledge, the development of an easy-to-use, rapid, non-invasive, and intelligent solution for (24/7) monitoring of MS, as well as its progression and medication effectiveness has not been proposed before. Along the same lines, the combination of simultaneous processing EEG, EMG, and electrooculography (EOG) for remote (at-home) and real-time monitoring of the patient’s brain and body activity before and after the medication has not been investigated. Visual evoked potentials (VEP) are specific brain patterns that are considered the only non-invasive way of showing in humans that there is demyelination or remyelination. VEP patterns are considered a key element to rapidly diagnosing and monitoring MS. En route to accelerating the adoption of new digital health technologies for neurological diseases diagnostics and monitoring, this research project’s goal is to develop and validate a novel hybrid brain-computer interface system (combining EEG, EMG, and EOG) to identify digital markers correlated with different MS stages (relapsing-remitting MS (RRMS), primary progressive MS (PPMS), and secondary progressive MS (SPMS), as well as propose a novel BCI system to rapidly monitor disease progression and treatment effectiveness. For that, EEG, EMG, and EOG will be recorded simultaneously using easy-to-use, wearable (gel-free), and certified (CE-mark class IIa & FDA-510k cleared) sensors provided by Myelin-H company.

Particularly, this research project will investigate the interpretation of evoked potentials when playing different games, particularly VEP, auditory evoked potentials, motor imagery movements, and Posner attention paradigms, which have been developed by Myelin-H as part of its clinical work on MS. Thereafter, novel processing and machine learning techniques (relying on spiking neural networks and neuromorphic computing) will be developed to decode collected data and translate them into powerful medical reports for neurologists to rapidly, remotely, and instantaneously assess treatment effectiveness and general MS patient’s health status. Developed algorithms will first be tested using retrospective data provided by Myelin-H and will thereafter be validated with different MS patients in Italy, the UK, Luxembourg, and Germany.

Applicant

Gullem Wissam Khairy

Project title

HEat transfer and propellant fluid phase transition investigation in WAter-based Space Propulsion systems (HEWASP)

Host institution

University of Luxembourg

Collaborating company

Bradford Deep Space industries (Bradford DSI)

Abstract

Back in the spotlight, the space market is in a great expansion and stretching the limits every day. New Space companies are constantly innovating and adapting their solutions to face the new challenges coming up along the space conquest. More specifically, deep space exploration arouses the interest of space agencies and spacecraft’s propulsion systems are an especially important source of innovations to break the current limits of space exploration and allow a self-sustainable economy in space. In response to those challenges, space companies are growing interest in electrical propulsion systems using water as propellant. Indeed, the use of water as propellant allows deep space satellites to refuel during their trip by mining water which is an abundant resource in space.

In this new space race, water-based propulsion enables safe, affordable, simple and sustainable operation of microsatellites. However, in order to build an efficient and reliable water-based propulsion system it is necessary to understand the physics occurring inside that system. Having more efficient propulsion systems is important to reduce their consumption of power, mass, and volume which are sensible limiting factors on spacecrafts. Inside an electric propulsion system, the liquid propellant passes through a heating chamber where it is vaporized and super-heated and gets finally expelled to create movement.

So, by understanding the physics happening to the water propellant in the specific conditions of the heating chamber, one can develop an accurate numerical model of a water rocket propulsion system. Then, having a numerical model gives to space engineers an effective way to develop and test new designs for the propulsion system with much less efforts than today. Thus, being able to predict the phenomena occurring during propulsion operations highly simplifies the design optimisation with the final goal to increase the performance of the propulsion systems which ultimately leads to an increase of the payload capacity of micro-satellites.

PhD – Sustainable Resources

Applicant

Haseeb Ur Rehman

Project title

A High-Resolution Numerical Weather Prediction Model for Nowcasting Precipitation in the Grand-Duchy of Luxembourg (NWPLux)

Host institution

University of Luxembourg

Collaborating company

RSS-Hydro S.a.r.l.(RH)

Abstract

Flash floods are one of the most dangerous and destructive natural disasters in the world. In Europe flash floods cause on average 50 casualties per year and overall, 70% of deaths due to natural hazards are associated with floods. Flash floods occur suddenly after high rainfall in a local area, are potentially highly destructive and are often considered as difficult to predict. During the last decade there have been several severe flood events in Luxembourg with the one in July 2021 being amongst the worst since records began.

The proposed project NWPLux is designed to make weather, particularly rain, forecasts in Luxembourg more accurate using a dedicated numerical weather model. Using this model, NWPLux will attempt to better predict where and when it is raining and use this information to make flood simulations. If successful, NWPLux could make a considerable contribution to a future flash flood warning system for Luxembourg, which could help to significantly improve flood preparedness and reduce damages.

To achieve this purpose, the proposed project will build the first dedicated weather model for Luxembourg that will generate highly accurate intensity and precipitation maps, which can then be used for flood simulations. This will allow organizations to make better flood predictions in shorter time. Besides global background weather data with low resolution, the NWPLux weather model will ingest local meteorological data which are collected from the MeteoLux station at Findel Airport, stations from the Administration de la gestion de l’eau (AGE) and from the Administration des services des techniques de l’agriculture (ASTA). In addition, it will also include the highly accurate and updated atmospheric water vapour data from Global Navigation Satellite System (GNSS), which is computed at the University of Luxembourg. This new weather model will then be calibrated considering past severe weather events that caused flash floods in Luxembourg and the output of this model will be used in a flood model.

Flood models allow to make simulations where and when water levels rise to flood levels, which can then be used to make warnings. It is important to note that NWPLux will use the same highly accurate height information for Luxembourg from the Administration du cadastre et de la topographie (ACT) in both the weather and flood models to improve accuracy in its results. The NWPLux would greatly benefit from the previous FNR projects PWVLUX and VAPOUR during which the methods were developed at the University of Luxembourg for the computation of atmospheric water vapour from GNSS. In parallel to NWPLux, the currently 12-station national GNSS network of the Satellite Positioning Service Luxembourg (SPSLUX) hosted by the ACT will be supplemented with extra low-cost GNSS stations by the NWPLux team in order to ensure a homogeneous and dense sampling of the atmospheric water vapour across the nation at no cost to the proposed project. NWPLux will also benefit from the recent FNR project SIPFLU and an associated PhD project, which both were performed in close collaboration between the University and with the private partner RSS-Hydro.

As the NWPLux team plans to work in close collaboration with AGE, ASTA, MeteoLux and the Flood Prediction Center Luxembourg, there will be benefits for all involved through knowledge exchange. Upon successful completion of NWPLux the benefit for society would arise from the potential of more accurate weather forecasts as well as improvements to the flood warning system. NWPLux provides the possibility to further national expertise in GNSS meteorology and NWP modelling while for the private partner, it would allow a strengthening of their related activities and services.

PhD – Law & Economics

Applicant

Krištof Horvat

Project title

Machine Learning-Based Price Modelling for Express Freight Network (PM-EF)

Host company

Eurosender SARL

Industry partner

University of Luxembourg

Abstract

Express freight is one of the crucial logistics solutions for companies to support just-in-time production and their needs for express deliveries. In case, some delay happens on the production line, express freight could solve the potential halt of the production lines. Eurosender is a digital logistics company acting as a broker on the logistics market. The company is sourcing the vehicle capacity from carriers that own vehicles and can perform the transport. On another side, clients can book a transport via the Eurosender website. As volumes are high, the company can offer affordable price solutions for clients that utilize Eurosender’s vast logistics network. Lack of clarity on the market is high as shippers and carriers have an incentive to hide the expected contracted price. Eurosender is a technology company with a mission to digitalize logistics processes. It is in an excellent position to optimize express freight transportation networks at scale in the European markets.

There are two possibilities to price transport on the logistics market; via contracts where the price is agreed between a carrier and a shipper in advance or on the spot market, where it is dynamically set for each lane and timeline differently. Currently, Eurosender has issues with setting a price on the market. On the spot market, the company does not utilize data to predict what should be the optimal spot price. Logistics agents set the price by feeling and not by the data.

This research aims to build a model that will calculate the best-selling price, based on different parameters, such as cargo type, distance, seasonality, et cetera. The second research objective is to investigate an alternative pricing solution called market-based pricing. It needs to be investigated whether Eurosender could introduce a market-based price instead of a contracted price.

The company has a growing issue with the load acceptance rate of the carriers as they often reject the load, even though the contract is in place. The research goal is to investigate which parameters influence the load rejections. To address issues explained in the previous paragraph, I aim to research the Eurosender past transactional data. To properly test different models, the implementation of research methods such as artificial neural networks and reinforcement learning will be needed. The research will be conducted within the Luxembourg Centre for Logistics and Supply Chain Management (LCL). The LCL is part of the MIT Global SCALE Network that offers an excellent opportunity for research collaboration.

To digitalize vital parts of the supply chain, an automatic supply and demand matching engine with more sophisticated pricing is an important part of building more robust transport networks. The study has a high implication on the pricing of freight networks, especially for the intermediaries’ broker companies that strive to improve supply and demand on the market.

Postdoc – Materials, Physics & Engineering

Applicant

Michal Habera

Project title

Constraint aware optimization of topology in Design-for-Additive-Manufacturing (COAT)

Host company

Rafinex S.à r.l.

Collaborating institution

University of Luxembourg

Abstract

Topology optimization is a modern method to find the ‘best’ shape for an engineering part. The ‘best’ means in this case that the shape will be stronger with the same amount of material used. Very often, the shapes obtained from this method cannot be easily or cheaply manufactured in large numbers. The reason is that processes and techniques used for manufacturing pose certain restrictions (constraints). For example, a 3D printer would not be able to print the letter “T” standing upwards because the ‘arms’ would fall off during printing. These types of problems must be respected by the shape to be produced, and it is not easy to find a shape that is both the ‘best’ one and can be cheaply manufactured.

This project aims to solve these challenges in two steps. In the first step, mathematical equations that force the topology optimization method to respect the constraints will be formulated. In the second step, these equations will be solved with the most advanced algorithms and computer software. None of the existing software currently allows this.

One of the outcomes of this project is to improve the existing open-source software to make this smart and efficient solution possible. New ideas from this project will find use in industry and academia. Rafinex will have the new functionality added to their product portfolio, while academic researchers working on topology optimization will have access to journal publications.

RELATED NEWS